IX
Focused Research and Development Needs

To reduce long-term risks at the sites and improve the basis for models and assumptions in the performance assessments, the committee recommends that the Department of Energy (DOE) carry out a focused research and development program to support tank closure activities. By “focused research and development activities” the committee means a program that concentrates on improving current technologies or developing technologies that could provide deployable results within 10 years. Although research needs concerning cementitious materials used in tank remediation applications, robotics, and chemical cleaning of tanks (from Chapters III, IV, and V) are discussed in this chapter, long-term research activities have been identified in the monitoring and performance assessment chapters (Chapters VI and VII), these would be carried out in parallel, but they are not the focus of this chapter.

Technologies that are deployable in 10 years could be developed and implemented during the tank remediation program and, in particular, in time to address the most challenging tanks (i.e., those with cooling coils, recalcitrant waste, or leaks), which are the ones most likely to have significant heels.1 The tank remediation program is a multi-decade endeavor and DOE has an opportunity to use this time to its advantage.

The committee believes that there are at least three critical topics warranting focused research and development efforts: (1) in-tank and downstream consequences of existing and advanced chemical cleaning options; (2) technologies to assist in tank waste removal, including robotic devices; and (3) near-term and long-term performance studies on those

These topics represent the greatest technological challenges (i.e., waste retrieval and tank cleanup) and knowledge gaps (i.e., long-term performance of cementitious material). In addition to the recognized technical challenges in the program, there may be some “unknown unknowns” suggesting additional technological vulnerabilities, that is, areas that warrant additional research and development that cannot be foreseen right now but may become apparent once DOE further progresses in its tank remediation program. DOE should undertake a systematic effort to identify the most important vulnerabilities to reducing programmatic and human health risk as one step to address these “unknown unknowns.”

The committee judges that a focused applied research and engineering development program aimed at reducing the amounts of waste left in the tanks or improving waste immobilization could lead to reduced risks on-site. Validating assumptions and improving DOE’s knowledge base could increase confidence in its waste management plans or its assumptions about long-term performance of the waste forms disposed of on-site, both of which are desirable outcomes. Moreover, these research and development activities could support the development of contingency approaches to address unanticipated difficulties in baseline processes. Research and development activities to address these topics are discussed below.

IN-TANK AND DOWNSTREAM CONSEQUENCES OF EXISTING AND ADVANCED CHEMICAL CLEANING OPTIONS

The sludge component of tank waste is a sticky, semi-solid material that forms from the agglomeration of oxides and hydroxides of iron, aluminum, and manganese and is a time-dependent consequence of the neutralization of nitric acid processing solutions with sodium hydroxide. Sludge

1

Previous National Research Council reports contain recommendations on long-term research and development needs for DOE’s Environmental Management Science Program and some specifically for high-level waste tanks (NRC, 1996a, 1999a, 1999b, 2000d, 2001a, 2001b, 2001c, 2001d, 2002b, 2003b). The committee recognizes the importance of an ongoing basic and applied research program to support DOE’s environmental management mission; however, such a program is not the focus of this chapter. cementitious materials used to fill tanks and immobilize low-activity waste.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 101
Tank Waste Retrieval, Processing, and On-Site Disposal at Three Department of Energy Sites: Final Report IX Focused Research and Development Needs To reduce long-term risks at the sites and improve the basis for models and assumptions in the performance assessments, the committee recommends that the Department of Energy (DOE) carry out a focused research and development program to support tank closure activities. By “focused research and development activities” the committee means a program that concentrates on improving current technologies or developing technologies that could provide deployable results within 10 years. Although research needs concerning cementitious materials used in tank remediation applications, robotics, and chemical cleaning of tanks (from Chapters III, IV, and V) are discussed in this chapter, long-term research activities have been identified in the monitoring and performance assessment chapters (Chapters VI and VII), these would be carried out in parallel, but they are not the focus of this chapter. Technologies that are deployable in 10 years could be developed and implemented during the tank remediation program and, in particular, in time to address the most challenging tanks (i.e., those with cooling coils, recalcitrant waste, or leaks), which are the ones most likely to have significant heels.1 The tank remediation program is a multi-decade endeavor and DOE has an opportunity to use this time to its advantage. The committee believes that there are at least three critical topics warranting focused research and development efforts: (1) in-tank and downstream consequences of existing and advanced chemical cleaning options; (2) technologies to assist in tank waste removal, including robotic devices; and (3) near-term and long-term performance studies on those These topics represent the greatest technological challenges (i.e., waste retrieval and tank cleanup) and knowledge gaps (i.e., long-term performance of cementitious material). In addition to the recognized technical challenges in the program, there may be some “unknown unknowns” suggesting additional technological vulnerabilities, that is, areas that warrant additional research and development that cannot be foreseen right now but may become apparent once DOE further progresses in its tank remediation program. DOE should undertake a systematic effort to identify the most important vulnerabilities to reducing programmatic and human health risk as one step to address these “unknown unknowns.” The committee judges that a focused applied research and engineering development program aimed at reducing the amounts of waste left in the tanks or improving waste immobilization could lead to reduced risks on-site. Validating assumptions and improving DOE’s knowledge base could increase confidence in its waste management plans or its assumptions about long-term performance of the waste forms disposed of on-site, both of which are desirable outcomes. Moreover, these research and development activities could support the development of contingency approaches to address unanticipated difficulties in baseline processes. Research and development activities to address these topics are discussed below. IN-TANK AND DOWNSTREAM CONSEQUENCES OF EXISTING AND ADVANCED CHEMICAL CLEANING OPTIONS The sludge component of tank waste is a sticky, semi-solid material that forms from the agglomeration of oxides and hydroxides of iron, aluminum, and manganese and is a time-dependent consequence of the neutralization of nitric acid processing solutions with sodium hydroxide. Sludge 1 Previous National Research Council reports contain recommendations on long-term research and development needs for DOE’s Environmental Management Science Program and some specifically for high-level waste tanks (NRC, 1996a, 1999a, 1999b, 2000d, 2001a, 2001b, 2001c, 2001d, 2002b, 2003b). The committee recognizes the importance of an ongoing basic and applied research program to support DOE’s environmental management mission; however, such a program is not the focus of this chapter. cementitious materials used to fill tanks and immobilize low-activity waste.

OCR for page 101
Tank Waste Retrieval, Processing, and On-Site Disposal at Three Department of Energy Sites: Final Report entrains varying amounts of insoluble actinide and fission products, and its presence therefore contributes to the overall source term of a particular tank. Most of the sludge waste can be mobilized and removed from storage tanks using traditional hydraulic techniques (i.e., mixer and transfer pumps; see Chapter III and Appendix F). Retrieval of residual sludge after the application of hydraulic techniques may require the application of chemical agents, aggressive sluicing methods, or a combination of both. Success with these approaches would further the inventory in, and potential future doses resulting from, individual tanks. As noted in Chapter III, DOE has demonstrated the efficacy of oxalic acid for the chemical cleaning of waste tanks in certain cases. Two of the three full-scale sludge dissolution trials were performed at the Savannah River Site (West, 1980; Fong, 1985; Adu-Wusu et al., 2003) and one at Hanford (Reddick, 2004). Oxalic acid cleaning was effective in the removal of additional sludge heel in Tank 16 at the Savannah River Site, moderately effective in Tank C-106 at Hanford, and ineffective for treatment or removal of zeolitic materials (used for radiocesium ion exchange) found in Tank 24 at the Savannah River Site. The effectiveness of oxalic acid for removing sludge residues derives from its ability to form stable, soluble oxalate complexes with the iron component of tank sludges. Oxalic acid cleaning was ineffective in the removal of additional sludge heel in Tank 24 at the Savannah River Site because zeolites are primarily aluminum silicates, and these compounds are more stable than aluminum oxalates—they are not “dissolved” by oxalic acid. Savannah River Site staff believes that oxalic acid has drawbacks associated with criticality safety, downstream processing, and costs (see Chapter III). However, the site recognizes the potential of chemical cleaning for treating sludge residuals in tanks with cooling coils where mechanical technologies for residual waste retrieval may not be effective. Savannah River Site staff has performed some limited research on alternative chemical cleaning agents and approaches to mitigating the potential adverse impact of oxalic acid. In a recent literature survey, Adu-Wusu et al. (2003) compared different chemical cleaning agents. The committee judges that chemical cleaning is a proven tank cleaning technology that could be effective in tanks with significant obstructions and therefore should be investigated further (see Recommendation IX-1). Two research and development paths for chemical cleaning can be explored: Cleaning agents other than oxalic acid that would not cause criticality concerns or downstream problems; and Methods to both predict and eliminate criticality concerns and downstream problems if oxalic acid is used as the cleaning agent. The degree of cleaning is coupled to assessing how much radioactive material is a reasonable amount to remain in the tank. A metric for assessing the need to remove the tank heel is a comparison of radioactivity in the tank to the radioactivity already committed to the site. At the Savannah River Site the total activity in the tanks is 426 MCi (1.58 × 1019 Bq), with a sludge activity estimated to be 203 MCi (7.5 × 1018 Bq; see Table II-1). The radioactivity remaining in the heel is estimated to be 2 percent of the total tank activity (USNRC, 1999),2 which would imply a heel activity of 8.5 MCi (3.15 × 1017 Bq). The heel isotopic composition will vary with the heel chemical composition. A heel composed of zeolites would be high in cesium-137. If the heel is primarily oxide precipitates, then strontium-90 and actinides would be the main radionuclides. At the Savannah River Site, around 11 MCi is already committed to the site; 18 KCi (6.70 × 1014 Bq) of transuranic waste and 11 MCi (4.07 × 1017 Bq) of low-level waste. The low-level waste has an isotopic composition that is different than the tank waste and includes a significant contribution from tritium. (The 11 MCi of low level waste is an overestimate since it is not decay corrected.) If the heels contain of 2 percent of the total tank radioactivity, then they will contribute a radioactivity burden to the Savannah River Site comparable to what is currently at the site. This type of assessment can help inform the determination whether additional tank cleaning is needed. Alternatives to Oxalic Acid The main problem with oxalic acid is the extremely large quantities that are used to neutralize the residual sludge and dissolve it: 26,000 to 38,000 kg of sodium oxalate per 5,000 gallons of sludge removed. Using a stronger acid, such as nitric acid (HNO3), to dissolve the sludge would reduce tremendously the amount of oxalic acid needed. Furthermore, nitric acid is an inorganic acid that does not complexate iron compounds (and, thus, does not raise criticality concerns in the waste) and also eliminates downstream problems such as foaming or CO2 releases that have been seen with oxalic acid. Because less chemical agent is used, the amount of secondary waste generated in the process is also smaller. Consequently, nitric acid would place a smaller burden on compliant tank space at the Savannah River Site, which is in 2 Another source (DOE, 2002) uses a value 15 times lower but the committee considers the retrieval estimates that underlie that value to be optimistic and unsupported (see Chapter VI). The values used here are estimates based on experience retrieving waste from Tank 16, a tank with cooling coils, before chemical cleaning (USNRC, 1999). DOE estimates that Tanks 18 and 19, which have zeolites but no coils, contain 28,000 Ci (1.0 × 1015 Bq) and 96,000 Ci (3.6 × 1015 Bq), respectively, after waste retrieval (Buice et al., 2005). These are 0.3 percent and 1 percent of the average total radioactivity per tank in tanks that have not undergone waste retrieval at the Savannah River Site, respectively.

OCR for page 101
Tank Waste Retrieval, Processing, and On-Site Disposal at Three Department of Energy Sites: Final Report scarce supply. Tank corrosion could be made relatively unimportant during tank cleaning because the duration of the chemical cleaning process is brief and the acid would be neutralized by the sludge. Any residual acid could be diluted with a rinse of inhibited water and/or neutralized with additions to the tanks, including the grout. Savannah River Site personnel indicated that they are in communication with other national laboratories, DOE sites, and Russian experts on tank cleaning technologies, particularly chemical cleaning. Westinghouse Savannah River Company, the main contractor in charge of tank chemistry, is already considering the use of nitric acid for residual sludge removal because of the critical tank space issue (see Appendix E and NRC, 2005a). A team was formed in December 2005 at the Savannah River Site to evaluate the use of nitric acid and its application to sludge removal in the Savannah River Site waste tanks. However, the site recognizes that the current flowsheet does not include nitric acid as a potential tank cleaning agent, so design safety analysis changes would be required if it were used. Other ligands can be used instead of oxalic acid to dislodge or dissolve the heel and remove the entrained radionuclides. Over the past 30 years, significant advances have been made in metal-specific complexation and ligand design that exploit metal ion speciation in producing selective complexes. A number of these advances are based on biomimetic studies that evaluate specific metal ion-ligand interactions in natural systems (Durbin et al., 1989; Raymond, 1990). Research has been performed on the leaching of actinides and fission elements from synthetic sludges under differing conditions and has been related to metal ion redox and speciation (Nash, 2002; Garnov et al., 2003). While metal-ligand interactions are understood primarily in the solution phase, ligands produced by bacteria (siderophores) will solubilize iron oxides, will dissolve oxides of uranium and plutonium, and can be exploited in developing metal-specific reactions for solid phases (Brainard et al., 1992). These metal-specific ligand approaches have been used for the selective removal of radionuclides from the human body (see e.g., Gorden et al., 2003). This same selective ligand approach has been applied to the area of radionuclide removal from tank waste (see e.g., Nash et al., 2000). To date, these advances have not been used in waste retrieval operations, but in the future they might offer cost-effective options for removing both solution- and solid-phase radionuclides from the tanks. Another approach to effect improved removal of entrained radionuclides from residual tank waste is to adjust the chemical oxidation state of the radionuclides. Metal ion solubility in a given aqueous phase varies dramatically with oxidation state. Plutonium is an excellent example with large differences in solubility that are dependent upon the metal ion oxidation state. This property is shared by other actinide elements and is exploited in nuclear fuel treatment for the dissolution of spent fuel and developing methods for tailored separations (Karraker et al., 2001; Thompson et al. 2002). Additional chemical cleaning agents may also be effective on recalcitrant waste types, such as zeolites at the Savannah River Site (and West Valley) and diatomaceous earths at Hanford. The committee has not identified any specific agents, but it is not clear that this possibility should be dismissed. Mitigation of Concerns Oxalic Acid The second possibility for research on chemical cleaning is to use oxalic acid while addressing the nuclear criticality concerns and downstream problems. The committee judges that the probability of a criticality event in a tank is low: It is unlikely that the tank waste processing system would either have a sufficient amount of fissile material in one location or configure it properly to start a chain reaction. Both of these conditions would be needed to achieve criticality. Based on the information provided by Savannah River Site staff on criticality concerns (see Chapter III), the committee was unable to determine whether the criticality concerns with oxalic acid are well founded. Therefore, the committee recommends continuing research on oxalic acid and carrying out a study that shows whether oxalic acid leads to criticality concerns in the tanks or downstream. Downstream problems could be addressed by destroying oxalic acid and metal oxalates after tank cleaning. The destruction of oxalic acid by oxidation has been investigated and can be used as the basis of further studies. Other oxidative methods have been investigated for the treatment of tank waste, including ozone, chemical oxidation (Patello et al., 1999), and electrochemical oxidation (Nash et al., 2003). While ozone is effective for destruction of oxalic acid in the laboratory, significant quantities would be required for these applications and care would need to be exercised to minimize occupational exposures. TECHNOLOGIES TO ASSIST IN TANK WASTE REMOVAL, INCLUDING ROBOTIC DEVICES DOE has been relatively successful in the limited number of waste retrievals undertaken to date. However, such success is tempered by DOE’s statement that it initially focused on retrieving waste from less complicated tanks. Future waste retrieval will become more difficult as more complicated situations (e.g., leaking tanks and tanks having substantial internal structures and more recalcitrant solids) are encountered. As a consequence, there is no assurance that previous successes will project into the future (i.e., that current retrieval technologies will be equally successful or even adequate). In part, DOE is addressing this issue by using a mix of available waste retrieval technologies modified to reflect specific circumstances and experience. As shown in

OCR for page 101
Tank Waste Retrieval, Processing, and On-Site Disposal at Three Department of Energy Sites: Final Report Chapter III, advanced waste retrieval technologies have been and continue to be developed by DOE or adapted from technology developed by others. DOE already acknowledges that future efforts to retrieve waste from its tanks will be challenged by the following: The environment within the tanks is one of high radiation and harsh chemical conditions. Retrieval is complicated by physical obstructions and recalcitrant chemical species. The waste characteristics relevant to retrieval are difficult to determine before retrieval because of the heterogeneous and anisotropic nature of the waste so DOE must “learn as it goes” on a tank-by-tank basis. The variation in the condition and contents of each tank makes its waste retrieval a unique undertaking. As shown in Chapter III, DOE faces the need to retrieve waste from many large tanks containing cooling coils and other obstructions, especially at the Savannah River Site. The baseline bulk retrieval approach consists of using water jets from the riser locations to spray material off the internal tank structures onto the bottom of the tank, consolidating it by sluicing, and pumping it from the tank. The potential limitations of bulk retrieval techniques are clear when one notes that the tanks containing a “jungle” of cooling coils are 23 to 26 m (75 to 85 feet) in diameter and up to 9 m (30 feet) in height. The efficacy of bulk retrieval in tanks with coils is uncertain: Good cleaning of tank surfaces appears to have been achieved in zones beneath the risers, but the amount of residual waste remaining in the “dead zones” between risers and in the tank periphery is unknown. Moreover, the limited accessibility of the tanks, the presence of cooling coils, and other obstructions severely limit the size and mobility of retrieval devices. With all of these challenges, bulk retrieval technologies are likely to leave significant amounts of waste in difficult environments that will challenge the capabilities of existing residual waste retrieval technologies. The committee judges that focused research and development investments in residual waste retrieval technologies suitable for tanks containing cooling coils or other obstructions, recalcitrant waste, or tanks that are leaking appear prudent. In Chapter III and Appendix G, the committee applauds the efforts at Hanford to develop the Mobile Retrieval System, a vacuum retrieval technology complemented by the in-tank vehicle, suitable for leaking tanks, as well as the development of the Salt Mantis, a high-pressure, low-volume water jet to mobilize recalcitrant waste deposits in tanks that have not leaked. Other potentially promising technologies (because they generate little or no secondary waste) include the use of high-pressure steam jets (used at the Savannah River Site to attempt to clean the annulus of Tank 16 and at Hanford in Tank C-106), CO2, and sodium carbonate pellet blasting (NRC, 2001c). The concept of an autonomous robot3 freely roaming inside a tank cleaning various surfaces to certain specifications with minimal human control has an intuitive appeal because of the potential for reduced labor costs and possibly avoiding complications posed by tethers. In general, since the 1970s, advances in robotic technology have been possible thanks to the progress of the microprocessor, although with a lag of some years. The continued increase of processing power, miniaturization, and speed has allowed robotic devices to make major advances in speed, precision, cost-effectiveness, scope of applications, and most critically, reliability. This is especially true in industrial applications that involve repetitive, predictable motions and tasks. More exotic applications such as space, service, military, and security robotics have also seen impressive gains commensurate with the funding levels invested in development (DARPA, 2005; NRC, 1996a).4 Because the challenges of DOE tank waste cleanup are unique and the opportunities for deployment have been few due to the pace of the tank waste cleanup program, development and deployment of robotic-like devices for this purpose has been attempted only by a few teams. DOE’s previous work and experience on articulated arms (e.g., the Modified Light Duty Utility Arm, in-tank vehicles such as Houdini and ITV at Hanford, and other examples cited in Chapter III and Appendix G) are worthwhile efforts in a necessary, continuing investigation on retrieval technologies. Most of these efforts were done within the Robotics Crosscutting Technology Development Program and Tanks Focus Area, which were both discontinued around 2002. The environment inside a tank, characterized by the following factors among others, is particularly hostile to untethered semiautonomous robotic technology: A potentially explosive, and of course radioactive, atmosphere; Sludge that impedes the mobility of robots and management of tethers; Physical obstructions such as cooling pipes and debris; Obstructed visibility for necessary vision systems through vapors, sludge, and physical obstructions; 3 For the scope of this report, a robot is a programmable, multitask manipulator capable of operating within a three-dimensional space and manipulating tools in response to its control programs; the robot may be tethered or untethered. To operate within the tank space the robot would befitted with on-board sensors whose information may modify its programs. Tele-operated articulated arms or tele-operated mobile vehicles (e.g., tank crawlers) are not robots unless they are programmed to perform their tasks autonomously. 4 For example, the Defense Advanced Research Projects Agency (DARPA) Grand Challenge is a series of races of autonomous robots through a desert course. Groups compete against each other to create the best autonomous robot that will complete a desert course avoiding all obstacles and following DARPA’s preset rules. The grand prize was set at $1 million in 2003 and $2 million in 2004 and 2005 (DARPA, 2005).

OCR for page 101
Tank Waste Retrieval, Processing, and On-Site Disposal at Three Department of Energy Sites: Final Report Contaminants hostile to instruments and sensors; Limited number and size of openings for accessibility (risers); Need for external power source for the robots through tethers or frequent recharging of batteries; and A challenging, unpredictable environment that is contrary to the programming benefits of robots and demanding of unique sensors and intelligence. The committee judges that the use of autonomous robots (i.e., operating without human control) for waste retrieval operations is not a realistic or practical alternative given the current status of the technology. Any waste retrieval tool for tank cleanup application must have some type of tether for the following reasons: It must be provided with some means to pull it out to safety in case repairs are needed; The tools carried are heavy and generate appreciable forces—hence the need for large actuators and appreciable power make the robot impractically bulky; without a battery, a tether would have to include an electric cable to supply power; and Tooling such as water jets requires a continuous feed. The management of a tether around pipes and through thick sludge is a challenging task. Moreover, the tank environment is not well defined, especially in tanks with vertical cooling coils, so robotics devices cannot currently be programmed to handle all possible situations that may be encountered. Given these challenges, developing a deployable and reliable, untethered, semiautonomous robot for waste retrieval purposes may take at least a decade of well-focused research, development, and deployment effort and investment on the order of $10 million per year. A previous National Research Council report (NRC, 2001a) suggests ideas for long-term research needs on untethered, semiautonomous robotic device. Robotic technologies will continue to advance, especially for the control of robots with sensors, and may accomplish in the future what is not possible today. Nonetheless, robotic technologies may still play a role in enhancing the effectiveness of available retrieval technologies. Existing waste technologies could benefit from “robotization” of some of their functions to simplify their operation, to reduce the level of skill required for their control, and to automate some aspects of the cleanup process. For example, mechanical arms and in-tank vehicles are practical platforms for deploying cleanup tools such as the Salt Mantis, the water mouse, and the wash ball. Robotic enhancements of articulated arms or in-tank vehicles could provide sensory feedback signals (e.g., force, temperature, visual signal) so that the operator can respond accordingly. Some operations, such as “go to point x,y,z or advance by 10 mm” could be programmed instead of manually jogging a tool precisely. Having a human in the loop to make operating decisions in response to sensory feedback signals is likely to prove most effective. The trade-off would be in balancing the cost of enhancements against the savings in cleanup time. For tanks with cooling coils, developments are warranted to enable the delivery of cleaning tools, such as high-pressure water jets, to surfaces shadowed by the cooling coils. For example, a directionally compliant tube could be pushed horizontally through the tank; this tube would bend sideways past the cooling coils without overstressing them and still deliver a water jet to the shadowed areas of the tank and pipes. Such enhancements may improve the rate of retrieval (i.e., reduce cost), but not necessarily increase the amount of waste removed. Other uses of robotics may help reduce exposure of workers to radiation. For example, commercially available industrial robots could be used to insert cleanup tools and retrieve them through tank risers, currently operations that can cause serious worker exposures. A targeted study may find similar uses for industrial robots to reduce worker exposure within the tank farms. Tele-operation with sensory feedback appears to be more amenable than programming to operate in the irregular environment of the tanks. NEAR-TERM AND LONG-TERM STUDIES REGARDING TANK FILL MATERIALS For the purposes of performance assessment, certain assumptions must be made about the short-and long-term behavior of cementitious materials used to fill tanks and immobilize low-activity waste. Their behavior with regard to mixing, pumping, and placement can readily be determined in short-term tests such as mockups and can be well informed by the experience of the construction industry. However, the same cannot be said for their long-term grout performance in service. When used for tank stabilization and low-activity waste immobilization for hundreds or thousands of years, cementitious materials are subjected to service conditions well beyond the experience of the construction industry. The ability to develop meaningful predictions of the behavior of cementitious grout or concrete over the long term requires a good understanding of fundamental mechanisms, such as Microstructure formation and degradation mechanisms; Pore solution chemistry (including pH and Eh); Binding properties (including toxic heavy metals and radionuclides, and also ions such as chlorides and alkalis that participate in deterioration processes);5 5 To some degree, the hydration products of cementitious materials are able to incorporate various foreign materials such as toxic heavy metals, thus rendering them immobile. However, how much and how effectively they can do so depends on the specific combination of cementitious materials and the pore solution chemistry. For example, slag is particularly effective at binding chloride ions, and in Portland cement the tricalcium aluminate component also binds chloride. However, if the pH of the pore solution is reduced—for example, by carbonation (reaction of the calcium hydroxide with carbon dioxide from the atmosphere)—the capacity of the hydration products to incorporate chlorides is reduced, and some of the bound chlorides will be released. Thus, the ability of the grout to bind foreign materials either incorporated in it at the time of mixing or migrating in from the outside can change over time.

OCR for page 101
Tank Waste Retrieval, Processing, and On-Site Disposal at Three Department of Energy Sites: Final Report Transport properties (permeability and diffusivity); and Mechanical properties. Much work has already been done on the science of concrete and other cementitious materials to advance our understanding of their microstructure, deterioration mechanisms, transport properties, and mechanical properties as they change over time. However, since most of this work has been done with construction applications in mind, little attention has been paid to their behavior decades after the initial mixing and placement. Also, with the exception of pH, relatively little work has been done with regard to changes in pore solution chemistry and binding properties because they have little relevance to the construction industry. Some work on the use of cementitious materials for the stabilization of toxic heavy metals has been performed and would be relevant to DOE applications. However there remain some gaps in our knowledge that are unlikely to be filled without DOE research and development. The committee is well aware that DOE is required by law to meet certain milestones in its progress toward closure of the tank farms. However, the committee judges that there is sufficient potential for better decisions about grout applications to warrant a research and development program. The committee believes this can be conducted without adversely affecting the overall schedule for closure of the tank farms. The following section describes the near-term focus of the proposed research and development program and then outlines its features. Near-Term (5-to-10-Year) Focus of Grout Research and Development The main focus of the research and development program in the near term would be to provide a sound basis for selecting the best formulations for grout based on the anticipated service environment and performance requirements. In keeping with the committee’s recommendation that DOE increase transparency and public involvement in decision making, this program would demonstrate that the grout formulations selected are clearly superior to the alternatives and would provide a quantifiable basis for comparing their performance. The concrete durability research conducted by Atomic Energy of Canada Limited for its near-surface disposal facility indicated that while service life predictions cannot be made with a high level of confidence based on a 5- to 10-year laboratory test program, such a program is helpful in providing a sound basis for comparing the performance of different grout formulations and thus for the selection of candidate grouts (Philipose, 1988; Feldman et al., 1989; Philipose et al., 1990a; 1990b). Although it is not possible to conduct real-time tests of grout durability for the relevant time frames, deterioration mechanisms can be estimated through accelerated testing. Such tests are performed under conditions that accelerate the degradation of concrete (e.g., elevated temperatures, electrical potentials to accelerate the migration of destructive ions, increased concentrations of destructive chemicals, cycling of temperatures, cycles of wetting and drying). Accelerated durability tests can be problematic however, because conditions imposed to accelerate deterioration may foster different mechanisms than would occur naturally. In addition, durable materials by definition take a long time to deteriorate, which necessitates long-term testing program to obtain results. To avoid these unrealistic test conditions, the proposed research and development program discussed below incorporates only methods that provide a modest degree of acceleration of the deterioration mechanisms, such as mildly elevated temperatures and increased concentrations of the chemicals of interest, similar to the test conditions at Atomic Energy of Canada Limited (Feldman et al. 1991; Philipose et al., 1991, 1992), and the recommended examination methods possess the ability to observe early signs of deterioration. Research and development on alternative grout placement technologies, such as jet grouting, to improve the degree of mixing of waste with grout could also yield results that are deployable in 5 to 10 years. Research and development work on jet grouting is already in progress at the Los Alamos National Laboratory (AEATES, 2004). Research on grout placement was performed at the three sites of concern in this report; Idaho National Laboratory conducted a mockup test that showing that a five-phase sequential pour would enhance the waste removal process (see Chapter V’s section on Tank Grouting at the Idaho National Laboratory). As mentioned in Chapter V, some short-term testing of grout materials has been conducted at DOE sites, and in particular, at the Savannah River Site. However, the committee has not seen any reports of long-term testing or a comprehensive analysis of basic properties to model long-term behavior. Experts at the Savannah River Site (Dr. C. Langton and Mr. T. Caldwell) reported that based on their tests, they have concluded there is effectively no mixing of grout with the insoluble tank heel, but the liquid is effectively incorporated into the grout or the dry cementitious materials deposited on top of the grout. Accordingly, DOE’s ongoing performance assessment does not take credit for mixing of the grout and heel but, as noted in Chapter VI, assumes that the grout maintains its structural integrity for 500 years and its physico-

OCR for page 101
Tank Waste Retrieval, Processing, and On-Site Disposal at Three Department of Energy Sites: Final Report chemical integrity for 10,000 years. The assumption concerning structural integrity is based on an earlier analysis done for the E Area Vaults performance assessment (Martin Marietta Energy Systems et al., 1994). It appears that the values used in the performance assessment for Tanks 18 and 19 for such properties as hydraulic conductivity have been inferred from the literature or from test data on the saltstone test program (see Chapter V). Understanding changes in these properties over time (decades, centuries, and millennia) is important to model waste releases to the environment in the performance assessment. Therefore, the committee cannot assess the assumptions of physical and chemical durability of the cementitious material used to fill the tanks and to immobilize low-activity waste. However, the committee believes that fundamental understanding of parameters that affect grout stability is not adequate to support the expected 500- to 10,000-year performance period. Outline of Proposed Grout Research and Development Program The U.S. Nuclear Regulatory Commission (USNRC) is cosponsoring research at the National Institute of Standards and Technology (NIST) on degradation mechanisms, mixing formulations, durability, and modeling of cementitious materials (Garboczi et al., 2005). DOE could provide the necessary input parameters for NIST’s microstructure-based model, such as The various environmental conditions present and anticipated at the three sites (e.g., soil chemistry, conditions of wetting and drying); The chemicals within the tanks or the low-activity waste; The ingredients for formulating grout available at each site (e.g., cementitious materials such as cement—ordinary Portland, sulfate resisting, and others), fly ashes, slag, silica fume; admixtures such as sodium thiosulfate (reducing agent); and A range of water-cementitious materials ratios. The DOE laboratories have the most intimate knowledge of the particular environment and performance requirements of the grouts and can provide essential insights into the best surrogates to simulate the chemical environment for safe testing. Universities bring a rigorous approach to research and often possess unique or specialized laboratory equipment. Industry has practical knowledge that can keep the program tied to the real world, as well as the capability to conduct a large-scale test program within the limits of budget and schedule while maintaining a high level of quality. A series of grout samples of different formulations could then be subjected to a succession of simulated environmental conditions over the duration of the program. Periodically, specimens can be taken out of the solution baths and “sacrificed” to locate the reaction front to determine the depth of ionic ingress into the specimens. Variation of depth of ingress with time for the various formulations is an indication of their transport properties and hence an indication of durability. Examples of examination methods would include the following: Petrographic examination. Petrography uses mainly optical microscopy, supplemented by scanning electron microscopy and sometimes chemical analyses or other methods. It is concerned primarily with determining spatial relationships among the hydration products of the cementitious materials and can identify early signs of changes in microstructure, the products of degradation reactions, and the mechanisms of degradation. The NIST model is based on microstructure; thus, the information provided by petrography is essential both to establish a starting point and to validate and refine the model’s predictions. An additional advantage of petrography is that the most informative specimens it employs are thin sections, which are slices of concrete approximately 20mm thick mounted on microscope slides. These specimens are extremely stable and remain usable for 100 years or more. Thus they could be archived to provide opportunities for examination and comparison many decades in the future. Such basic information would be an asset to both model validation and the site’s long-term monitoring efforts discussed in Chapter VII. Bulk chemical analysis. Where the mechanism of interest is the ingress of some specie, a series of thin slices of grout can be analyzed (e.g., with an electron microprobe) for that chemical to determine concentration gradients, from which the transport properties of the chemical can be determined. A study of the changes in the characteristics of leachate over time would inform the performance assessment. In this example, monitoring would continue beyond 10 years to allow comparisons with and updating of the performance assessment to increase confidence that the grout is working as designed.6 More basic research activities that could be performed in the same 5- to 10-year time frame include identifying and evaluating oxidation pathways and kinetic mechanisms for grout degradation. Some of these activities could also be conducted in parallel with saltstone to compare retention capabilities for the mobile radionuclides such as technetium-99. Many of these research and development needs were 6 The tanks or vaults and cementitious materials within them would have to degrade before water could get in, leach contaminants, and leak out of them. The time at which this might occur is uncertain, and there may not be any water leaching out of these tanks for decades. Hence, this monitoring activity would have to be coordinated with the long-term monitoring activity (see Chapter VII).

OCR for page 101
Tank Waste Retrieval, Processing, and On-Site Disposal at Three Department of Energy Sites: Final Report identified by Westinghouse Savannah River Company at the committee’s meetings. USE OF TEST BEDS FOR THE STUDY OF RETRIEVAL TECHNIQUES As noted in Chapter III, waste retrieval (bulk or residual) has not yet occurred in most tanks at DOE sites. Many of the tanks contain internal features such as cooling coils or other debris that promise to impede waste retrieval. In these situations, new technology or adaptations of existing technology may be desired or required. Adapting an existing retrieval technology or deploying a new retrieval technology in a radioactive environment can cost millions of dollars, and failures can cost even more. Thus, it is technically prudent and cost-efficient to test retrieval technologies in non-radioactive test facilities (test beds) before attempting to deploy them. The usual approach to ensuring that a radioactive waste process will work consists of two steps: tests with the actual material (hot tests) at the laboratory scale to ensure that the process fundamentals are understood (e.g., sludge dissolution, pumpability), and tests with nonradioactive simulants (cold tests) at large or full scale to prove the design and the equipment.7 The purpose of a test bed is to: Test all equipment prior to field deployment, Provide hands-on experience with retrieval systems under simulated conditions, Troubleshoot retrieval system design and operations, Train equipment operators; Support development of operating procedures and maintenance protocols, and Test off-normal (outside of planned operation or behavior) and recovery activities. In 2002, the Hanford Site built a test bed called the Cold Test Facility (CTF) at a cost of $2.87 million (Dodd, 2005). The CTF consists of a 75-foot diameter open-top tank to simulate a single-shell tank or a double-shell tank. The CTF can hold up to 600,000 gallons of simulated waste. A steel bridge or primary superstructure spans the open tank to accommodate full-scale mixer pumps and transfer pump system mockups along with waste retrieval equipment. Although the structure is open, it is capable of simulating customized constraints such as hanging interference mockups and risers pits. Tests are sometimes conducted at night to simulate the lack of visibility inside the tank. Retrieval technologies demonstrated at the CTF include the in-tank vehicle crawler, an off-riser sampler, and the hydrolaser water lance (Salt Mantis), and technologies used at other sites (see Table F-2 and Appendix G). The Savannah River Site has operated a test bed at the TNX facility, but it is not clear that this test bed will be available in the future. The Pump Test Tank is a partial Type IV tank mockup at the mostly decommissioned TNX facility, used for testing and equipment before deployment. The Idaho National Laboratory does not have a test bed on its facility, but it conducted two major mockup activities: (1) to test grout placement techniques (see Chapter V); and (2) to test waste retrieval equipment (see Chapter III). The first mockup facility was constructed in 1999 at an Idaho Falls industrial facility. It consisted of a full-scale horizontal slice of the bottom of a tank only a few feet in height, but 50 feet in diameter, with cooling coil structures and simulated residual solids. Grout mixtures were tested to validate assumptions of flowability in both the tank and the vault areas and the ability of the grout to move the in-tank solids toward the steam jet (INEEL, 1999). The second mockup facility was constructed and used during 2000-2001 at another Idaho Falls industrial facility. This mockup was used to test the ability of the spray wash system to clean solids from the tanks. A full-height tank mockup was constructed, with only half the circumference. Simulated solids were placed in the tank, and the spray system components were operated to validate assumptions of their cleaning ability. In addition, assumptions regarding the ability of the steam jets were tested to ensure that heavy solids loadings could be removed from the tanks without plugging the steam jets (INEEL, 2001). The benefits of test beds have long been recognized by DOE staff and contractors. Often cited among waste retrieval lessons learned is the importance of a cold test facility to test all equipment before deployment (Caldwell, 2005a; Dodd, 2005; Lockie, 2005a). The following is a quote from a Pacific Northwest National Laboratory review of lessons learned in tank waste remediation at Oak Ridge National Laboratory (Bamberger et al., 2001; p. 7.1): First and foremost, cold testing is extremely beneficial prior to any first-time field deployment. Not only does this initial testing in a clean environment allow any significant design flaws to be identified and reworked before contamination controls become a significant issue, but it also provides valuable training for the operators and craft personnel by providing them with an opportunity to become familiar with the equipment from the inside out. In addition, integrated cold testing allows development of procedures that reflect how operations are actually conducted and allows multiple operators to receive training under low pressure conditions rather than “on the front lines.” Finally, cold testing can provide important opportunities to demonstrate readiness as part of a phased readiness review process. 7 Laboratory-scale tests and tests with simulants did not reveal the difficulties that emerged when DOE used the in-tank precipitation process to remove cesium from waste in Tank 48. That process, however, is rather different in nature than the type of waste removal and tank cleanup technologies that the committee describes here.

OCR for page 101
Tank Waste Retrieval, Processing, and On-Site Disposal at Three Department of Energy Sites: Final Report The committee recommends that one or more radioactive test beds for retrieval technologies that can be adapted to simulate a variety of tank situations (e.g., recalcitrant heels, cooling coils, debris) be maintained and made available for other DOE sites. LEVEL OF RESEARCH AND DEVELOPMENT INVESTMENT Savannah River Site staff is collaborating with DOE’s Office of Cleanup Technologies to develop technologies for the following: Chemical cleaning methods for heel removal The impact of oxalic and other acids on carbon steel tanks during heel removal Selective dissolution of targeted radionuclides during heel removal Mechanical methods for heel removal that minimize water usage Removal of fast-settling solids (e.g., zeolites) and hardened sludge Efficient removal of salt heels Criticality control during heel removal Characterizing and mapping residual material after heel removal Cleaning methods for tank annular space Methods for verifying waste removal from tank annular space Another example of a technology development initiative is the tank cleaning technology exchange meeting that is scheduled for the first quarter of 2006. The purpose of this meeting is to share technology development and deployment efforts among DOE sites and vendors to identify equipment and systems applicable to tank closure that have been successful within DOE and in commercial applications. New technologies that do not require as much high-level waste tank space are also being pursued (Ross, 2006b). The committee recommends a centralized, focused, 10-year bench-scale research and development program with a budget amounting to a minimum of $10 million per year and, more desirably, $50 million per year to support DOE’s tank cleanup program. This is based on experience with the Environmental Management Science Program (EMSP; prior to its transfer to DOE-Office of Science), the Office of Science and Technology Tanks Focus Area (TFA), and Robotics Crosscutting Technology Development Program, and similar programs such as, Defense Advanced Research Program Agency (DARPA) programs. The magnitude of these investments is justified as follows. The initial EMSP budget was approximately $50 million per year at the beginning in 1996 and decreased to about $30 million in 2003 when the program was moved to the DOE-Office of Science. The average size of EMSP grants is about $500,000 for three years.8 In 2000 there were 306 research projects, of which 76 were categorized as high-level waste problem areas (i.e., relevant to the tanks), which represented at yearly budget of $38 million. The Tanks Focus Area’s yearly budgets9 before it was disbanded were $28.5 million (FY 1997), $30.1 million (FY 1998), $29.1 million (FY 1999), $47.9 million (FY 2000), $41.5 million (FY 2001), and $43.125 million (FY 2002). In addition to bench-scale research effort, new technologies may require studies using real waste (hot tests) and pilot-scale studies using test beds (e.g., tank mockups equipped with coils and surrogate heels) to test their effectiveness before full-scale deployment. Hot and pilot-scale studies will require additional funds, realistically $50 million per year and possibly more. A budget of $50 million per year would correspond to approximately 3 percent of the FY 2007 DOE-Environmental Management (EM) budget request for tank waste storage, retrieval, treatment, immobilization, and disposal at the three sites (roughly $1.6 billion) (OMB, 2006).10 A higher research and development investment as a percentage of the program’s budget would bring the program more in line with research and development investment in industry and other federal agencies.11 A $50 million per year budget for 10 years ($500 million total) also represents one percent or less of DOE’s estimated cost for the tank cleanup 8 Projects and budgets are available on the EMSP web site at: http://emsp.em.doe.gov. 9 These numbers do not include the management costs to support the Tanks Focus Area activities. 10 The President’s 2007 proposed budget requests $104.5 million for radioactive liquid tank waste stabilization and disposition at Idaho National Laboratory, $571 million for tank waste stabilization and disposition at the Savannah River Site, and $964 million for Hanford’s Office of River Protection (which is responsible for the storage, retrieval, treatment, immobilization, and disposal of tank waste). The proposed budget for research on high-level waste treatment and storage is zero dollars (termination of all research), and the proposed budget for technology development in DOE-Environmental Management (for which tank wastes are listed as the top priority) is approximately $21 million (OMB, 2006). 11 A 2001 report by the National Research Council found that research and development efforts in other federal agencies are 9 percent for EPA, 15 percent for DOD, and 40 percent total for DOE, but only 4 percent for Environmental Management (NRC, 2001d). Although the research and development budgets in private industry are not directly comparable to research and development in the federal government, the research and development intensity (research and development funding as a percentage of net sales) “is highest [12-13.7 percent] in knowledge-intensive industries such as software and pharmaceuticals, whereas research intensity is lowest [0.66-0.73 percent] for such mature industries as petroleum and construction” (NRC, 2001d). The report goes on to conclude that DOE’s environmental quality mission is fairly knowledge intensive. It is beyond the scope of this present study to update the statistics from six years ago, but these numbers do give a sense of the scale of research and development investment that other organizations deem appropriate. DOE’s tank waste mission is arguably not as knowledge intensive as the software industry, but the tank waste program needs research and development more to fulfill its mission than do the mature industries cited.

OCR for page 101
Tank Waste Retrieval, Processing, and On-Site Disposal at Three Department of Energy Sites: Final Report program ($50 to $75 billion12). The research and development program would fund applied research and engineering development projects in the high-priority areas identified in this chapter and those identified through an assessment of the engineering vulnerabilities in DOE’s tank remediation program (i.e., cases in which performance poorer than planned causes significant problems for the program). Projects should be selected using a competitive process similar to that used for the EMSP and would be in addition to or in place of any directed (noncompetitive) awards. The tank remediation program at the sites should be responsible for supporting field testing and deployment of technologies resulting from this program. However, program management should be centralized to ensure prioritization and coordination of the projects. The Tanks Focus Area showed the effectiveness of centralizing research and development activities at DOE tank sites. Despite variations in site and tank characteristics, many technological issues were common among two or more sites. The research and development needs outlined in this chapter and elsewhere in the report are common to all three sites so it seems more cost-effective to develop common solutions for the sites, where appropriate. According to the last Tanks Focus Area report, the application of technical solutions from the program accounted for more than $250 million in cost savings (or avoidance) with a projected life-cycle savings of more than $5 billion (DOE-TFA, 2001). The Tanks Focus Area was also an effective platform to centralize and coordinate results from other components of DOE’s Office of Science and Technology, such as the Environmental Management Science Program; DOE’s Environmental Management Accelerated Site Technology Deployment Program; crosscutting programs (e.g., efficient separations and processing, robotics, characterization monitoring, sensor technology); and DOE-Environmental Management’s cooperation with industrial partners, universities, and national laboratories. Between the program’s inception in 1996 and its disbandment in 2002, the Tanks Focus Area studied more than 200 technology applications that resulted in approximately 160 deployments, including the wash ball used to clean the interior of the tanks at the Idaho National Laboratory, the caustic-side solvent extraction proposed for the Salt Waste Processing Facility at the Savannah River Site, and the Light-Duty Utility Arm and the Houdini in-tank vehicle used for waste retrieval at the Idaho National Laboratory. Numerous other technologies were demonstrated at DOE sites and showed promise for implementation, including saltcake dissolution, residual waste mapping, and sampling technologies (DOE-TFA, 2001). FINDINGS AND RECOMMENDATIONS Finding IX-1: Based on experience with the Environmental Management Science Program (prior to its transfer to DOE-Office of Science), the Office of Science and Technology, Tanks Focus Area, Robotics Crosscutting Technology Development Program, and similar programs such as the Defense Advanced Research Program Agency programs, a centralized, focused, 10-year bench-scale research and development program with a budget amounting to a minimum of $10 million per year, and more desirably, $50 million per year seems reasonable, if pilot-scale studies are included. Recommendation IX-1: DOE should fund applied research and engineering development projects in the high-priority areas identified in this chapter and those identified by the vulnerability analysis. Projects should be selected using a competitive process similar to that used for the DOE Environmental Management Science Program. These projects and their funding would be in addition to or in place of directed (noncompetitive) awards, if any. The tank remediation program at the sites should be responsible for supporting field testing and deployment of technologies resulting from this program. However, the program management should be centralized to ensure prioritization and coordination of the projects. Finding IX-2: Oxalic acid has proven to be an effective chemical cleaning technology in Tank 16 at the Savannah River Site and, to a certain extent, in Tank C-106 at Hanford. The committee believes that oxalic acid can be a helpful cleaning tool in tanks with significant obstructions. There are signs of interest within DOE or its contractors at the Savannah River Site, but currently there is no active research and development on chemical cleaning. Recommendation IX-2: DOE should fund research and development partnerships among universities, national laboratories, and industry focused on options for chemical cleaning of tanks to find alternative cleaning agents or to mitigate the criticality and downstream processing problems that Savannah River Site staff pointed out to the committee. Finding IX-3: Untethered semiautonomous robotic devices are not likely to add value to the retrieval process in tanks, given the current state of the technology. Proven technologies, such as an articulated arm with a water jet and remotely controlled vehicles with pusher blades, provide for retrieval tool deployment and visual feedback to human operators and 12 In 2003, DOE estimated that by implementing its plan for accelerated cleanup, the department could reduce the projected $105 billion cost and 70-year time frame for cleanup of tank wastes at the Savannah River Site, Hanford, and Idaho National Laboratory to $76 billion and 35 to 50 years. The General Accounting Office (now the Government Accountability Office) found opportunities for further cost and schedule savings and errors in DOE’s estimates on the order of billions of dollars (GAO, 2003). Yet even with these corrections and allowing for uncertainties, the tank waste cleanup program is expected to cost at least $70 billion over several decades.

OCR for page 101
Tank Waste Retrieval, Processing, and On-Site Disposal at Three Department of Energy Sites: Final Report represent the limit of useful retrieval technology at this time. Robotic enhancements to proven human-controlled techniques are likely to yield more effective performance than untethered, semiautonomous robotic devices. Recommendation IX-3: During tank cleanup and closure operations, DOE should continue the investigation and development of more effective technologies for waste retrieval from tanks with emphasis on tanks with obstructions and recalcitrant waste. In particular, DOE should continue to use, adapt, and improve the most effective suite of available waste retrieval technologies on a tank-by-tank basis, including consideration of technologies at other DOE sites, in industry, and internationally. Additionally, if retrieval beyond the capabilities of existing physical methods is required, DOE should Reevaluate the use of oxalic acid to determine whether criticality and downstream problems are a real concern; If they are a real concern, DOE should investigate methods to mitigate these shortcomings or develop alternative acceptable chemical cleaning approaches; and If chemical cleaning proves to be inapplicable to some types of recalcitrant deposits (such as zeolites), DOE should develop alternative mechanical methods, such as delivery tools and techniques for waste mobilization that enhance waste accessibility to water jets. Finding IX-4: In the near term, decisions about the formulation of grouts for tank fill and immobilizing low-activity waste are being made on the basis of experience in very different applications and in some cases also on data from short-term tests on saltstone. The committee has not seen any reports of long-term testing or of more fundamental research directed at the unique aspects of DOE applications, particularly the binding capacity of grouts and the changes in various properties over the extended times contemplated by the DOE. Recommendation IX-4: DOE should initiate a focused research and development program over a 5- to 10-year period, and longer where necessary, to improve the fundamental understanding of the long-term performance of cementitious material and to tailor different formulations of grout to different tanks or groups of tanks, and different low-activity waste compositions. The program should involve collaboration among government laboratories, universities, and industry. Finding IX-5: In addition to the research and development areas identified in Findings 1-3, future research and development needs may become apparent as DOE progresses in its tank cleanup program. Recommendation IX-5: DOE should support an independent assessment of its tank remediation program for the purpose of comprehensively identifying and prioritizing any additional technical vulnerabilities in the program as a basis for funding additional research and development. The vulnerability assessment should be independent of, but rely heavily on, information obtained from the tank remediation programs at the sites. Finding IX-6: The benefits of test beds have long been recognized by DOE staff and contractors at sites where waste tanks need remediation. Often cited among waste retrieval lessons learned is the importance of a cold test facility to test all equipment before deployment. Recommendation IX-6: DOE should maintain one or more radioactive test beds for retrieval technologies that can be adapted to simulate a variety of tank situations (e.g., recalcitrant heels, cooling coils, debris) and make them available for use by other DOE sites.