Advice on the Department of Energy's Cleanup Technology Roadmap: Gaps and Bridges (2009)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

2 Principal Science and Technology Gaps The first part of the statement of task for this study requests that the committee identify principal science and technology gaps and their priori- ties for the cleanup program. Previous National Research Council (NRC) reports have identified science and technology shortcomings using a variety of terms, for example, research needs, technology needs, cleanup chal- lenges, and knowledge gaps (NRC 2007c). To address its task statement, the committee first sought an informative definition of the word “gap” (Figure 2.1). The word “gap” is defined as a “discontinuity between two points” and in this context the task of identifying gaps could be interpreted to mean that the committee is to identify “showstoppers,” that is, cleanup tasks for which there is insufficient knowledge or technology available to do the task. Information provided to the committee by the Office of En- vironmental Management (EM) and its contractors indicated that, if suf- ficient time and money were available to overcome cleanup obstacles, there are no showstopper gaps in the cleanup program. Another way of stating this is that EM and its contractors are confident that technologies EM has incorporated into its cleanup plans and schedules (baseline technologies) can be made to work. Nevertheless, the committee observed in its interim report (Appen- dix H) that the complexity and magnitude of EM’s cleanup task requires the results from a significant, ongoing R&D program if EM is to complete its cleanup mission safely, cost-effectively, and expeditiously. To address its statement of task in a way it judged would be most useful to EM, the committee chose as its working definition that a gap is a shortfall in avail- 21 

22 ADVICE ON THE DOE’S CLEANUP TECHNOLOGY ROADMAP FIGURE 2.1  The word “gap” can be seen as the intersection of several synonyms. Source: Visual Thesaurus: http://www.visualthesaurus.com/. 2-1 Bitmapped able knowledge or technology that could prevent EM from accomplishing a cleanup task on its expected schedule and/or budget. Following the anal- ogy of a roadmap, a science and technology gap is a “pothole” in the road that EM might somehow work around, but at the likely cost of time and money. It would be much better to fill the pothole or avoid it altogether with appropriate research and development (R&D). Addressing potholes could help EM to avoid large, insurmountable problems by addressing smaller technological challenges that could other- wise aggregate into showstoppers. Smaller investments in developing new science and technology could allow funding for several R&D approaches to a pothole problem, which would be more likely to lead to the most effective solution. Filling potholes before they erode into washouts is a natural role for EM’s longer-term roadmapped research. GAP IDENTIFICATION AND PRIORITIZATION Gap identification began with the committee’s March 2007 workshop and review of earlier Academies’ reports (NRC 2007c), and then proceeded through the committee’s site visits, which are summarized in the appendixes of this report. In this chapter, gaps are set forth as problems or potholes, 

PRINCIPAL SCIENCE AND TECHNOLOGY GAPS 23 which, if avoided or fixed with new technical tools, could help make the EM cleanup safer, faster, or less expensive. The identification of a gap does not imply that a given baseline technology might not work or should be abandoned—rather the gaps are incentives for EM to apply R&D to improve its available site cleanup and remediation tools. The titles of the gaps are intended to factually state a situation or condition that is less than optimal. The gaps have been identified at a level that the committee believes will provide EM the insights and flexibility to develop and implement effec- tive, bounded, and targeted R&D to fill the gap. EM’s draft Engineering and Technology Roadmap, issued in April 2007, provided a framework for organizing the committee’s fact-finding and deliberations, but the committee worked independently of the specific contents on the Roadmap. Factors qualitatively considered when identify- ing the gaps included: • Whether the gap required medium- to long-term R&D, • The volume of waste affected, • Potential to reduce technical risks (including risk to workers), • Reduction in schedule uncertainty, • Potential cost savings, • Likelihood of a successful outcome to the R&D effort, and • Possible existence of solutions outside EM. Applying these criteria to information received by the committee led to the general list of about 50 science and technology issues given in Appendix C. Later, through the course of its deliberations, the committee refined this list to the set of 13 principal gaps described in this chapter. In the committee’s judgment, each of these 13 principal gaps could affect the schedule, cost, and risk associated with the EM cleanup program. The priorities of the principal gaps were determined by the commit- tee through an iterative process. During the August 2008 closed session, committee members who initially drafted gap analyses described the attri- butes of each gap to an “investment committee” composed of three com- mittee members who previously held major programmatic and budgetary responsibilities. The three suggested initial gap priorities according to the  Essential features of the EM Science and Technology Roadmap are outlined in Chapter 1, Sidebar 1.2. For convenience it will be referred to as the EM roadmap or simply as the Roadmap throughout this report.  These time frames are described in Sidebar 1.3.  The three members of the investment committee were Carolyn Huntoon, former Department of Energy (DOE) Assistant Secretary for Environmental Management; Edwin Przybylowicz, former vice president for research at Eastman Kodak; and Andrew Sessler, former director of Lawrence Berkeley National Laboratory (LBNL). 

24 ADVICE ON THE DOE’S CLEANUP TECHNOLOGY ROADMAP information presented. The full committee then refined the gap analyses, prioritization criteria, and priorities. Gaps were prioritized as high, me- dium, or low within each program area (Table 2.1). The committee did not attempt to prioritize the gaps across the pro- gram areas because the program areas differ fundamentally in the nature of TABLE 2.1  Principal Science and Technology Gaps and Their R&D Priorities Gap Numbera Statement of Gap Priority Roadmap Program Area: Waste Processing WP-1 Substantial amounts of waste may be left in tanks/bins after their High cleanout—especially in tanks with obstructions, compromised integrity, or associated piping. WP-2 Low-activity streams from tank waste processing could contain Medium substantial amounts of radionuclides. WP-3 New facility designs, processes, and operations usually rely on Medium pilot-scale testing with simulated rather than actual wastes. WP-4 Increased vitrification capacity may be needed to meet schedule High requirements of EM’s high-level waste programs. WP-5 The baseline tank waste vitrification process significantly Medium increases the volume of high-level waste to be disposed. WP-6 A variety of wastes and nuclear materials do not yet have a Low disposition path. Roadmap Program Area: Groundwater and Soil Remediation GS-1 The behavior of contaminants in the subsurface is poorly High understood. GS-2 Site and contaminant source characteristics may limit Medium the usefulness of EM’s baseline subsurface remediation technologies. GS-3 The long-term performance of trench caps, liners, and reactive Medium barriers cannot be assessed with current knowledge. GS-4 The long-term ability of cementitious materials to isolate wastes High is not demonstrated. Roadmap Program Area: Facility Deactivation and Decommissioning DD-1 D&D work relies on manual labor for building characterization, High equipment removal, and dismantlement. DD-2 Personal protective equipment tends to be heavy and hot and Low limits movement of workers. DD-3 Removing contamination from building walls, other surfaces, and Medium equipment can be slow and ineffective. aReferred to throughout this report. 

PRINCIPAL SCIENCE AND TECHNOLOGY GAPS 25 the risks that R&D could mitigate and in their timescales. Establishing their relative priorities involves policy judgments that are outside the committee’s expertise. These differences and trade-offs are elaborated briefly below. 1. Cleanup work that involves only human activities, such as facil- ity construction or demolition, can be accomplished on human-controlled schedules. Groundwater and soil remediation, on the other hand, involve geologic processes that humans can only attempt to control and, gener- ally speaking, operate on a much longer timescale. Priorities for R&D, which typically include schedules and expected payoff, will be different for “engineering-only” projects versus those involving geologic processes (Sidebar 2.1). 2. The ultimate goal of site cleanup is to protect humans and the SIDEBAR 2.1 Timescales for Engineering Projects Versus Geologic Processes A conceptual incongruity exists in the time domains for DOE site cleanup between those activities associated with tank closure, waste separation and processing, and demolition and decommissioning and those activities associated with groundwater and soil contamination, environmental remediation, and long- term site stewardship. The pace of activities associated with the former is limited primarily by budget resources, which are manifested as physical infrastructure, size of the labor force, and availability of chemical and engineering solutions. Public and regulatory policies also modulate the selection and implementation schedules, such as tank closure. On the other hand, activities associated with soil and groundwater protection and cleanup are substantially controlled on a geologic scale by natural process rates and characteristics such as permeability of aquifers and the vadose zone, and the massive volumes of contaminated material (albeit at much lower concentration of problematic toxic organics, metals, and radionu- clides than that of the original source materials). Simply put, removing waste from a tank is a straightforward, although complex, engineering challenge that can be addressed as such; soil and groundwater remediation cannot be similarly planned and accomplished. A billion-dollar investment in tank closure or in a waste processing facility is likely to have a dramatic effect within a few years, while a like investment in groundwater remediation may only marginally accelerate the schedule for site cleanup and closure. The end state for the groundwater remediation at a site may be ultimately determined by an acknowledgment and acceptance that the site cannot be returned to a pristine state but that the residual contamination is suf- ficiently well understood scientifically that the risk to the public, end users, or the environment is acceptable, or can be reduced or controlled at acceptable levels far into the future. 

26 ADVICE ON THE DOE’S CLEANUP TECHNOLOGY ROADMAP environment from long-term consequences of the nation’s former produc- tion of materials for nuclear weapons. Essentially this means remediating contaminated groundwaters and soils and protecting them from future contamination, which would argue that the higher priorities be given to the groundwater and soil gaps. However, the successful recovery and pro- cessing of waste from the former production operations, especially the high-level tank waste, is a prerequisite for preventing additional releases of contamination to the environment—in both the near and long terms. Facil- ity decontamination and demolition, no matter how carefully it is planned and conducted, carries the potential for immediate injury or fatality among workers—as opposed to possible future consequence from groundwater and soil contamination. The program areas are thus intertwined and none rises to a higher priority than the others. SCIENCE AND TECHNOLOGY GAP ANALYSES The science and technology gaps that are presented in the remainder of this chapter are arranged according to the main program areas in the EM roadmap (DOE 2007a, 2008b). Each gap is analyzed in terms of how it is an obstacle for EM, and R&D opportunities to deal with it are described, as follow: • An overview of the nature of the gap, • The impact the gap has on EM’s cleanup program, • The current status of work related to the gap, and • Future R&D approaches that EM could consider to help bridge the gap. A table at the end of each gap analysis shows the basis for the com- mittee’s assessment of the gap’s priority. Factors qualitatively assessed for each gap were volume of waste affected, potential to reduce technical uncer- tainty, potential to affect cleanup schedule, and potential to affect cost. To illustrate how these factors were evaluated, the millions of gallons of high- level tank waste ranked as high in the volume category, as did contaminated groundwater. R&D for gaps that reflected lack of knowledge tended to rate high for reducing technical uncertainty. Schedule and cost reductions are not always correlated, for example, for groundwater and soil remediation or waste treatment processes that are already under way. WASTE PROCESSING The waste processing program area of the EM roadmap deals primarily with high-level tank waste issues, including waste storage, waste retrieval, 

PRINCIPAL SCIENCE AND TECHNOLOGY GAPS 27 tank closure, waste pretreatment, and waste stabilization. Millions of gal- lons of high-level waste (HLW) from reprocessing nuclear fuels to recover plutonium and other nuclear materials arose during the Cold War era. Hanford, the first site that reprocessed fuels on an industrial scale, used several different reprocessing technologies that resulted in a variety of waste compositions. The Savannah River Site (SRS) mainly used one type of reprocessing technology and has a relatively smaller spectrum of waste compositions. Reprocessing activities at Idaho were at about one-tenth the scale of those at Hanford or SRS. There were important similarities and differences in waste management practices among these three sites, which are reflected in the science and technology gaps in waste processing identi- fied by the committee and described in this section. Waste Processing Gap 1 (WP-1): Substantial amounts of waste may be left in tanks/bins after their cleanout—especially in tanks with obstructions, compromised integrity, or associated piping. Waste from former weapons material production at the Hanford site (Appendix D) is stored onsite in 149 single-shell (single-walled) and 28 double-shell tanks. The single-shell tanks were constructed between 1943 and 1964. The last of the double-shell tanks was constructed in 1986. All of the double-shell tanks have capacities of 1 million gallons. In total, 133 of the tanks have capacities of 500,000 to 1 million gallons. The Hanford tanks currently hold about 54 million gallons of waste, which contain a total of about 193 million curies of radioactivity (NRC 2006b). Hanford tank waste is very heterogeneous, but generally speaking it consists of su- pernatant liquid, water-soluble salt cake, and insoluble sludge. These phases resulted from the original acidic reprocessing waste being made alkaline for compatibility with the waste tanks, which were built from carbon steel, and subsequent evaporation of water to reduce the waste volume. The phases are layered and intermixed to varying degrees. Sludge removal is the most difficult. SRS has 49 tanks in service that hold about 36 million gallons of waste containing about 426 million curies of radioactivity (Appendix G). The SRS tanks have a variety of designs—some single-shell, some double-shell, and some with the secondary shell less than the full height of the primary tank (i.e., “cup in a saucer”). The tanks vary in capacity from 750,000 to 1.3 million gallons. Most of the SRS tanks have internal cooling coils that were used to keep the temperature of the waste below boiling (NRC 2006b). SRS tank waste is broadly similar to that at Hanford although it is less hetero- geneous chemically (Figure 2.2). Most of the highly radioactive waste at the Idaho National Laboratory (INL) site is in the form of granular solids, which are stored in sets of stain- 

28 ADVICE ON THE DOE’S CLEANUP TECHNOLOGY ROADMAP FIGURE 2.2  Sludge sampled from an SRS tank. Tank sludge was formed by neu- tralizing acidic waste from the reprocessing of nuclear fuels and recovery of nuclear materials. This sludge flowed like a thick paste. Other sludges are more viscous or nearly solid, which makes them difficult to remove from the tanks. This approxi- 2-2 new mately 2-liter sample was opened inside a shielded laboratory cell like that shown in Figure 3.2 in the late 1970s. SOURCE: Department of Energy. less steel bins contained in concrete vaults (Appendix E). The calcine waste exhibits a variety of sizes and compositions. It was originally transferred pneumatically into bins for storage, and DOE plans to retrieve the calcine essentially the same way. However, pneumatic retrieval could be difficult if the calcine has caked (e.g., from moisture in the bins or by particle-to- particle sintering). According to a presentation to the committee during its site visit, INL used a simulated calcine to demonstrate technical approaches for removing the binned calcine (Hagers 2007). The INL site still has about 900,000 gallons of acidic liquid waste stored in three stainless steel under- ground tanks. According to the Ronald Reagan National Defense Authorization Act of 2005, Section 3116, for the tanks and their associated piping at SRS and Idaho to be closed, waste must be removed as much as is practical and 

PRINCIPAL SCIENCE AND TECHNOLOGY GAPS 29 meet the performance objectives in 10 CFR 61.40. Closing a tank marks the end of EM cleanup activities for that tank. Tank closure is EM’s top priority for site cleanup, and it is a top priority among public citizens and their representatives. The criteria under which Hanford tanks can be closed has not yet been established (NRC 2006b, Johnson 2008). Most of the legacy waste tanks at Oak Ridge were closed years ago although a few small surge and collection tanks remain (Appendix F). Impact of the Gap Tanks containing waste heels that have not been removed to the “ex- tent practical” according to the Reagan Act or that cannot be shown to meet specified performance objectives to limit long-term radiation exposure cannot be closed. Tanks containing appreciable amounts of residual waste (heels) are unlikely to be accepted by DOE, its regulators, or the public for closure. Removal of the bulk of the waste with large pumps (for SRS and Han- ford) or pneumatic devices (for INL) appears to be relatively straightfor- ward and efficient. However, experience at Hanford and SRS has shown that sludge heels inevitably remain in the tanks after the bulk of the waste has been retrieved. Reducing the volume of this heel becomes increas- ingly difficult, time-consuming, and expensive as the volume of the heel declines. The tanks at Hanford and SRS generally have small access ports ­(risers); some tanks contain debris, and at SRS cooling coils further inhibit access and waste retrieval (Figure 2.3). A number of single-shell tanks at Hanford have leaked waste into the environment, and some double-shell tanks at SRS have leaked waste into the annulus between the tank walls (Figure 2.4). The structural integrity of tanks that have leaked is considered to be com- promised. Buried waste transfer lines and ancillary equipment (e.g., smaller tanks, valves, transfer pits, and pumps) also contain waste.  If these criteria are met, DOE can designate the residual waste as “waste incidental to reprocessing,” a legal distinction that allows it to be permanently disposed onsite. Otherwise, classified as HLW, it would have to be removed for disposal in a licensed repository such as that proposed at Yucca Mountain, Nevada.  Actions to close a tank after the waste removal criteria are met include isolating it from the waste system and filling it with a material such as grout with no intent for further waste retrieval. 

30 ADVICE ON THE DOE’S CLEANUP TECHNOLOGY ROADMAP FIGURE 2.3  Cooling coils in an SRS tank. Such coils maintained the temperature of high-level radioactive waste below boiling. The coils are an obstacle to removing the tank waste at SRS. SOURCE: Department of Energy. 2-3 new Current Status Oak Ridge completed cleaning eight concrete-walled tanks in 2001, and all together closed 65 tanks between 1995 and 2007 (NRC 2007c). These tanks are smaller than those described above, but nonetheless demon- strated the use of several types of innovative remotely operated equipment, which led to substantial savings in cost and schedule (Boyd 2008). At the time of the committee’s visit, Hanford had retrieved the waste from seven single-shell tanks, and waste retrievals were in progress or planned for four others (Mauss 2007). SRS has closed two tanks and is expected to have four more ready for closure by 2010. None of these tanks had internal cooling coils or other significant obstructions. The cleaning of a tank annulus has not been attempted. Both Hanford and SRS operate tank mock-ups in which waste-retrieval challenges are simulated and new technologies are tested. In 2005 an EM subcontractor successfully retrieved simulated calcine from a bin (AEATES 

PRINCIPAL SCIENCE AND TECHNOLOGY GAPS 31 FIGURE 2.4  Salt accumulated in a tank annulus. Double-walled tank construction helped to prevent the release of radioactive waste into the environment. In this figure minor leaks from the primary wall (right) have accumulated in the annulus. SRNL 2-4 new recently developed a robotic crawler for cleaning the tank wall. SOURCE: Department of Energy. 2005). The sites have little experience in removing waste from bins, tank annuli, transfer pipes, or ancillary equipment. Approaches to Bridge the Gap Residual waste retrieval from tanks and ancillary pipelines was identi- fied as an important technology gap in three NRC reports (2001b, 2003, 2006b). These reports recommended the development of physical and chemical cleaning technologies to improve the effectiveness of residual waste removal in tanks, tank annuli, and pipelines, especially technologies that reduce the risks of leakage of wastes to the environment during the removal operations (e.g., by using little or no water to retrieve wastes). Opportunities for expanding the use of robotics technologies for waste re- trieval and tank cleaning are discussed in NRC (2006b). Site presentations at Hanford (Honeyman 2007) and SRS (Davis 2008; Spears 2008) included a number of technology needs for improving waste retrieval (Appendixes D and G). 

32 ADVICE ON THE DOE’S CLEANUP TECHNOLOGY ROADMAP According to this committee’s assessment of information it received, the following approaches have promise for future EM R&D: 1. Chemical approaches that do not degrade the tanks or cause down- stream problems, but that can dissolve recalcitrant (or agglomerated) sol- ids in nonflowing areas (e.g., behind cooling coils, in clogged pipes). This could include R&D to mitigate the downstream effects of known chemical approaches. More extensive knowledge in the areas of (i) structure and dynamics of the materials and interfaces of relevance to the waste tank chemistry, (ii) complex solution-phase phenomena, and (iii) coupled chemi- cal and physical processes might lead to transformational engineering solu- tions to this problem, 2. More autonomous physical approaches (e.g., focused water jets, grinders, pushers) to break up agglomerated waste and remove waste from surfaces while minimizing water use, 3. Faster and more autonomous physical approaches to corral solid materials (e.g., pushers) and to remove them from the tank while keeping water volumes low, and 4. Efficient approaches to demolish and remove internal tank structures to allow access for waste retrieval and reduce water intrusion pathways. The sense of this committee, as well as the previous Academies’ com- mittees that considered EM’s challenges with tank residues (NRC 2001b, 2006b), is that the sites need a variety of technologies—a toolbox—that can be applied on an as-needed basis to maximize EM’s ability to retrieve its variety of waste types under a variety of tank conditions. Prioritization of the Gap Relative to other science and technology gaps discussed in this section the committee prioritized this gap as High. Relative Rating Criteria High Medium Low Volume of waste affected X Potential to reduce technical uncertainty X Potential to affect cleanup schedule X Potential to affect cost X 

PRINCIPAL SCIENCE AND TECHNOLOGY GAPS 33 WP-2: Low-activity streams from tank waste processing could contain substantial amounts of radionuclides. The separation of the tank waste into high-activity and low-activity streams is a key to EM’s plans for dealing with the waste (NRC 2006b). The sludge in Hanford and SRS tanks contains most of the waste radionuclides but comprises only about 10 percent of the tank waste volume at each site. The tank waste supernate and salt cake comprise about 90 percent of the waste volume (WP-1 and Appendixes D and G). Disposing of the salt cake and supernate (collectively called “salt waste”) as HLW would increase HLW treatment and disposal costs—which are already the largest single component of the EM cleanup—about 10-fold. The salt waste, however, contains most of the radiocesium (Cs-137) and small but significant amounts of Sr-90, Tc-99, and transuranic elements (TRU) such as plutonium. Accordingly, the supernate and dissolved salt cake will be processed before being stabilized (“pretreated”) to remove key radionuclides from this soluble waste and route the nuclides into the HLW stream. At SRS the low-activity waste (LAW) stream is to be incorporated into a cementitious grout, referred to as saltstone, for permanent onsite disposal. SRS encountered a significant obstacle in the mid-1990s when its planned salt processing technology was halted due to safety concerns. After several years of R&D, EM selected a new technology called “caustic-side solvent extraction” (CSSX) for SRS salt processing (Moyer et al. 2005). Development of CSSX began at Oak Ridge National Laboratory (ORNL), and it has continued with EM support at other national laboratories, espe- cially the Savannah River National Laboratory (SRNL). The key new CSSX facility—the Salt Waste Processing Facility (SWPF)—is scheduled to begin operating in 2013 (Appendix G). Because its waste tanks are filled nearly to capacity, SRS has developed interim measures for processing its salt waste in order to free up tank space until the SWPF becomes available. One of these, the modular caustic-side solvent extraction unit (MCU), is essentially a pilot-scale test of CSSX with actual tank waste. SRS operates its liquid waste facilities under State of South Carolina permits that impose a limit of 1.4 million curies in saltstone. Operation of the MCU and another interim process, referred to as deliquification, disso- lution, and adjustment, will put about 1.2 million curies into the saltstone, while pretreating only a small fraction of the salt waste (Appendix G). Thus DOE has a very high expectation for the performance of the SWPF—to add only 0.2 million curies of radioactivity to the saltstone while processing  Salt waste is primarily sodium nitrate, sodium nitrite, and sodium hydroxide. 

34 ADVICE ON THE DOE’S CLEANUP TECHNOLOGY ROADMAP the vast majority of the SRS salt waste. This appears to be a significant technical risk. The difficulties at SRS suggest that Hanford, with its more diverse tank waste compositions and characteristics, is likely to face challenges in processing its liquid wastes. Hanford has selected a newly developed ion exchange resin to remove radiocesium. While there has been extensive development work on this resin, the process has not been demonstrated in actual production. Impact of the Gap The SRS tank closure program cannot proceed without the ability to meet radionuclide separation objectives for its salt waste. Salt processing methods and objectives for the Hanford Waste Treatment Plant (WTP) are still in development. Baseline technologies are developed for both sites, but neither site has demonstrated the expected separations in actual plant op- erations. If the expected high separation factors are not achieved in actual operations, substantial program delays would be likely. There has been no decision as to whether Idaho calcine will require processing. Work in Progress In addition to developmental work on CSSX, SRNL has also made considerable progress in improving sorbents to be used to remove strontium and TRU from salt waste. These radionuclides contribute much less radio- activity to the salt waste than does cesium, but unless removed they could prevent the salt from meeting criteria that allow it to be disposed onsite. Along with the MCU, SRS is operating an actinide removal process (ARP) for its interim salt processing. The ARP will become a “front end” to the SWPF. Both Hanford and SRS are seeking higher-performance sorbents for salt processing (SRNL 2007; Tamosaitis 2007). SRS is also considering a technology called small column ion exchange, which uses the ion exchange resin being developed at Hanford, to acceler- ate its salt processing. Ion exchange columns would be inserted into risers in a waste tank and fed from the tank itself—saving much of the cost of constructing a processing facility. Another recent technology development, the rotary microfilter, would be used to prepare a solids-free feed to the columns. This process was described as being near technical maturity (Da- vis 2008).  It is not clear if radioactive decay is included in these numbers. Because of its 30-year half-life, about half of the Cs-137 in today’s waste inventory will have decayed by the end of the SRS site cleanup program. 

PRINCIPAL SCIENCE AND TECHNOLOGY GAPS 35 Approaches to Bridge the Gap Baseline separation processes and materials may not be as efficient or tolerant of impurities or process upsets as expected. Based on this com- mittee’s assessment of information it received, the following have promise for future EM R&D toward making LAW processing more robust and efficient: • Solvent extraction processes are susceptible to performance deg- radation from impurities. For example, silica, which is recycled in waste from borosilicate glass production in the Defense Waste Processing Facility (DWPF), can lead to poorly separable emulsions. Improved understanding of CSSX from operation of the MCU processes would be useful in increas- ing the efficiency of the MCU as well as ensuring that options are found for any operational issues that might arise. This, in turn, could help ensure that the SWPF meets its stringent performance requirements as discussed earlier. • Organic ion exchange resins for cesium removal can be regener- ated and reused, but are more readily degraded by radiation and corrosive chemicals than inorganic ion exchange material. Current inorganic ion exchange material generally cannot be regenerated and thus could contrib- ute significantly to the waste volume. R&D to improve the lifetime of the organic resins or develop an elutable inorganic resin would address these problems and significantly enhance the efficiency of pretreatment. Prioritization of the Gap Relative to other science and technology gaps discussed in this section the committee prioritized this gap as Medium. Relative Rating Criteria High Medium Low Volume of waste affected X Potential to reduce technical uncertainty X Potential to affect cleanup schedule X Potential to affect cost X WP-3: New facility designs, processes, and operations usually rely on pilot-scale testing with simulated rather than actual wastes. EM and its contractors are challenged with designing, building, and operating large, expensive, one-of-a-kind waste processing facilities that have major inherent safety risks because of the nature of the waste to be 

36 ADVICE ON THE DOE’S CLEANUP TECHNOLOGY ROADMAP processed. The Hanford WTP and SRS SWPF are examples. In addition to being highly radioactive and chemically hazardous, the wastes exhibit a wide range of chemical and physical properties. Developing waste treatment processes and scaling them up with assur- ance that they will work in production with actual waste feed streams is difficult. Working with actual waste in laboratory hot cells is generally on the gram to kilogram scale, which may not yield accurate data for process scale-up, and is cumbersome and slow. R&D at the pilot scale using actual radioactive materials can be exceedingly expensive—requiring engineering and construction work similar to building the full-scale facility itself—and therefore usually not feasible. To avoid these pitfalls R&D normally proceeds along two complemen- tary paths: bench-scale hot cell work with radioactive materials and pilot- scale work using nonradioactive simulants. However, developing a simulant that accurately represents the characteristics of each radioactive waste composition can take significant effort, especially for sludges. Because the composition of the sludge in any given tank may vary significantly, and different simulants may be required for different aspects of a single sludge (e.g., rheological properties, chemical properties, radiolytic properties), multiple simulants can be required for a particular waste stream. The committee heard of numerous basic waste processing operations that carry significant technical risk because they cannot be tested on a pilot- plant scale with actual wastes, including: • Reliable separation of solids from liquid waste streams to prevent clogging of ion exchange beds or adverse effects on solvent extraction equipment, • Ensuring that shear-thickening (non-Newtonian) sludges can be transported in pipelines without clogging, • Predicting the rate of radiolytic hydrogen generation by process sludges and the release time and rate of the hydrogen, • Predicting the stability and interaction of various process streams to allow for reduced conservatism in the operational safety bases of the tank farms, and • Understanding the effect of impurities and degradation or corro- sion products on process performance. Impact of the Gap Failure to accurately predict the flow properties of slurries is a leading cause of process failure (Merrow 2000). The absence of adequate under- standing of the behavior of process streams can necessitate overly conser- vative and costly process designs to minimize the risk of a process failure 

PRINCIPAL SCIENCE AND TECHNOLOGY GAPS 37 or the risk of unrecognized safety issues, which as a worst case can render a facility inoperable with the actual radioactive waste it was intended to process. An overly conservative process flowsheet can prevent efficient opera- tion of tank farms (e.g., prevent some wastes from being combined or processed), which increases costs and the time required for tank cleanup. The use of trial-and-error methods to develop representative simulants is costly and tends to increase reliance on even more costly hot cell operations. Additionally, many simulants contain hazardous materials that lead to high disposal costs for the simulants and equipment contaminated with them. Work in Progress Pacific Northwest National Laboratory (PNNL) is building a quarter- scale engineering “platform” to test and demonstrate WTP pretreatment operations using simulants—including sludge washing, leaching, and waste concentration (Figure 2.5). PNNL also has an extensive program to test pulse jet mixers, which are key components for mixing solid/liquid slurries in several WTP operations, with simulants. PNNL considers the liquid-solid mixing problems in design of the WTP to be on the forefront of mixing science (Michener 2007). EM has sponsored hot cell R&D to improve understanding of the in-process behavior of key radioactive materials at all four of the national laboratories visited by the committee. EM, in conjunction with the national laboratories, has begun to use lab- and pilot-scale data to verify and calibrate computational fluid dynamic (CFD) or other types of numerical models. These verified models can then be exercised through the range of conditions that might be encountered by varying parameters (e.g., flow rate, viscosity, solids loading, gas loading) to see what the effect on equipment operation is. Based on this information, the process design can be modified as necessary. The use of computer mod- eling verified through engineering tests is the mainstay of most industrial organizations and especially industrial chemical producers. Approaches to Bridge the Gap According to this committee’s assessment of information it received, computer modeling (e.g., CFD) can help bridge the gap between data that can be obtained from lab-scale tests with actual wastes and pilot-scale tests  A worst case happens rarely. However, one example is the SRS in-tank precipitation pro- cess for radiocesium removal, which behaved unexpectedly during a full-scale test with actual radioactive waste and was abandoned (NRC 2000a, 2001a, and Appendix G). 

38 ADVICE ON THE DOE’S CLEANUP TECHNOLOGY ROADMAP FIGURE 2.5  Diagram of pretreatment semiworks for the Hanford WTP. Such pi- lot-scale semiworks are essential for testing new processes before they are used to treat actual radioactive waste in large, expensive new facilities. Semiworks testing is done almost exclusively with nonradioactive simulants. The operators depicted on the right indicate the size of this semiworks. SOURCE: Department of Energy. with simulants. However, the use of computer modeling to replace large pilot- and full-scale testing with simulants carries some technical risk. These technical risks could be reduced if CFD or other models of rela- tively complex behaviors could be calibrated using data from tests with actual wastes. The models would then be used to predict the fluid system’s behavior under other conditions. Engineering tests under those conditions would determine the degree to which the computer-generated predictions were met. This approach could be used for a number of different phenom- ena including heat transfer, fluid flow in tanks and porous media, explosive atmosphere testing, chemisorption phenomena on resins and other solid media, and precipitate formation in heat exchangers and on pipes, pumps, and vessels. An essential component of bridging the gaps among waste simulants, computer models, and the behavior of actual waste will be R&D aimed at discovering potential, unexpected interactions or other phenomena inherent in the actual wastes that could lead to a process upset or failure. Such dis- covery-oriented R&D would help ensure that the conceptual model, which is manifested by the computer model, is correct. 

PRINCIPAL SCIENCE AND TECHNOLOGY GAPS 39 Prioritization of the Gap Relative to other science and technology gaps discussed in this section the committee prioritized this gap as Medium. Relative Rating Criteria High Medium Low Volume of waste affected X Potential to reduce technical uncertainty X Potential to affect cleanup schedule Xa Potential to affect cost X aAddressing this gap would significantly impact only new construction in the complex. It could also provide technical support or lead to modification of facilities that are finished or well along in construction and design, such as the SWPF and WTP. WP-4: Increased vitrification capacity may be needed to meet schedule requirements of EM’s high-level waste programs. A Joule-heated melter is being used at SRS to stabilize high-activity tank waste in borosilicate glass. Hanford is taking a similar approach, but the Hanford WTP is being designed to vitrify both high-activity and low-activity streams from its tanks using different melters for each stream. Joule-heated melters might be used at Idaho depending on future decisions concerning the disposition of calcine. The design and operation of Joule-heated melters for vitrifying tank waste limits their throughput. A slurry of waste and water is added to a “cold cap” on top of the molten glass. Water evaporates from the slurry, forming more of the cold cap. Material at the bottom of the cold cap is gradually incorporated into the melt. Heat to produce the glass comes from passing electricity directly through the melt (i.e., the Joule effect). Mixing within the melter is convection-driven and inherently slow in the viscous melt. New approaches such as bubblers are planned for use at the WTP to aid in mixing the melt and increase throughput. Although bubblers can increase the rates by up to 50 percent, other operational issues arise such as increases in the amount of volatilized chemicals and radionuclides that would have to be trapped in the off-gas treatment system. Even with bub- blers, the WTP would have the capacity to vitrify only about one-third of Hanford’s LAW. This is driving Hanford’s plan to use a supplemental treatment process for the additional LAW. If additional capacity can be obtained in the current space allocated for the WTP melters, then the need for supplemental facilities could potentially be avoided. 

40 ADVICE ON THE DOE’S CLEANUP TECHNOLOGY ROADMAP Impact of the Gap The throughput of a Joule-heated melter is relatively low and increasing the size is not practical since the physical installations are already complete. The number of years that the WTP and DWPF will operate depends on the time it takes to vitrify the waste on the respective sites. With each addi- tional year of operation costing about $500 million in today’s money (Davis 2008), increasing the melter throughput by a factor of 2 could potentially save several billion dollars on each site. Tamosaitis (2007) described several factors, in addition to the large volume of waste, that make WTP throughput a concern. These include the diversity of waste input streams, behavior of solids in the system, and process upsets. He listed improved waste forms, glass formulations, and melters as technology needs for enhancing throughput. At SRS both Spears (2008) and Davis (2008) listed increasing the throughput of the DWPF as technology needs. Davis stated that options needing further R&D include improving the glass-forming frit, improving the ability to mix the contents of the melter, and operating at a higher tem- perature with an alternative melter design. Work in Progress Hot-wall induction melters are in use in France and the United King- dom, and R&D on cold-wall induction melters is being performed in France and Russia (Ahearne 2002). INL researchers showed the committee a high- throughput induction melter that is being used for R&D with simulated wastes (Appendix E). Similar work has been performed at SRS based on operations at France’s Marcoule site (Barnes et al. 2008). Plasma-based melters that were capable of high throughputs within a small footprint were also examined by Westinghouse (McLaughlin et al. 1994, 1995). In addition, other R&D such as the bubbler work at the Catholic University of America and Russian work using microwaves to help heat the cold cap of the melter (Kurkumeli et al. 1992) have been carried out. Considerable R&D on alternative stabilization technologies for Hanford LAW, including bulk vitrification and steam reforming, is also ongoing at INL and Han- ford. NRC (1999a) recommended that DOE examine a range of technical options for immobilizing HLW calcine at Idaho if the calcine itself is not adequate to meet final disposal requirements. 

PRINCIPAL SCIENCE AND TECHNOLOGY GAPS 41 Future Approaches to Bridge the Gap SRS has a current need to increase the capacity of the existing DWPF melter to meet current programmatic needs. Simply installing a bigger melter is precluded by the design of the DWPF. At Hanford more capacity is needed, especially for the large volume of LAW. The WTP at Hanford is in the advanced design stage. Based on this committee’s assessment of the information it received, the following have promise for future EM R&D: • Alternative melter designs, with special attention to induction melt- ers or other types of melters that have high throughput relative to their size. • New methods of enhancing mixing in Joule-heated melters to in- crease their capacity. Bubblers can improve mixing, but they have potential technical problems associated with corrosion/erosion of the bubbler tubes and other components, and disturbing the cold cap. • Alternatives for boosting heat input to Joule-heated melters. Mi- crowave heating might be one option (NRC 2005). • New glass frit formulations that have lower viscosities to allow improved convective heat transfer in Joule melters, for example, adding Li2O to the frit. Any new formulation will have to be tolerant of variations in the waste stream. Prioritization of the Gap Relative to other science and technology gaps discussed in this section the committee prioritized this gap as High. Relative Rating Criteria High Medium Low Volume of waste affected X Potential to reduce technical uncertainty X Potential to affect cleanup schedule X Potential to affect cost X  Disturbing the cold cap would increase the load of volatile materials and radionuclides on the melter off-gas system or possibly trap water beneath portions of the melt, which could lead to eruptions of steam and molten glass. 

42 ADVICE ON THE DOE’S CLEANUP TECHNOLOGY ROADMAP WP-5: The baseline tank waste vitrification process significantly increases the volume of high-level waste to be disposed. Along with the rate at which a Joule-heated melter can produce the borosilicate glass waste form discussed in WP-4, the percent of waste that can be incorporated into the glass determines the rate at which tank waste can be processed (vitrified) for disposal. There are a variety of factors that determine the amount of waste that can be incorporated into a given vol- ume of glass (waste loading). These factors include the temperature of the melt, the composition of the glass-forming material (frit), and the composi- tion of the waste. Tank waste sludge contains some constituents, such as sulfate (SO42–), phosphate (PO43–), and chromium, that are relatively insoluble in the glass melt. The waste also contains aluminum, which increases the viscosity of the melt, making it hard to pour from the melter, and sodium, which r ­ educes the durability of the glass. Generally speaking, the best sludge load- ings that are currently expected under production conditions with currently available vitrification flowsheets are in the range of about 30 to 40 weight percent dry sludge with 60 to 70 weight percent of the added frit. In other words, current HLW forms are made from about one-third waste and two- thirds binder. Impact of the Gap Coupled with the limited throughput of current melters, the low waste- to-glass ratio establishes a decades-long time frame for working off the high-activity waste inventories at Hanford and SRS. Tank operations and a good deal of site infrastructure will have to remain open and operating to support waste vitrification during these decades. The approximately $500 million per year cost at each site for maintaining its high-activity waste operations is a strong incentive for faster waste processing (Appendixes D and G). A prolonged waste processing schedule also increases the chance that some tanks storing the waste, which have already exceeded their design lives, will leak. Current Status There are two approaches for increasing glass waste loading that are being investigated at Hanford, PNNL, SRS, and SRNL. One is to modify the glass formulation so that more waste can be incorporated into the glass matrix without compromising its processability or quality of the product. EM-supported work on improved frit formulations has allowed the waste 

PRINCIPAL SCIENCE AND TECHNOLOGY GAPS 43 loading of glass produced by the SRS DWPF to be increased from roughly 30 percent to around 38 percent. The second approach is to pretreat the waste to remove bulk nonradio- active constituents (aluminum, iron, sodium) to reduce the waste volume to be vitrified or to remove waste constituents that have low solubility in borosilicate glass (chromium, sulfate) and thus limit waste loading. R&D concerning selective removal of nonradioactive, solubility-limiting con- stituents from the sludge is being pursued at SRS and Hanford. Removal approaches have focused on water washing to remove sodium salts and washing with caustic solutions to remove aluminum and chromium. Future Approaches to Bridge the Gap According to this committee’s assessment of information it received, the following have promise for future EM R&D: • Additional work on understanding the chemical nature of nonra- dioactive components in high-activity waste that might lead to their selec- tive removal. Some forms of aluminum can be removed readily by caustic washing of the sludge, which has been demonstrated at SRS. Removal of more recalcitrant forms of aluminum may be necessary for aluminum re- moval to become a practical way of reducing waste volume, especially at Hanford where tank conditions (heat, age) have probably produced more of the recalcitrant forms. Chromium in Hanford waste will reduce waste loadings in WTP glass unless it is removed in pretreatment or reduced in concentration by blending with low-chromium waste. • Work to develop entirely new, nonborosilicate glass waste forms that can accommodate higher waste loadings and/or loadings of problem- atic constituents like aluminum, chromium, and sulfate. Phosphate glasses were one alternative class of waste forms described to the committee. • Waste forms that include little or no added binder. Idaho calcine is one such example. Perhaps sintered or minimally bonded sludges could be developed for Hanford and SRS. Such work would probably rely heavily on computer modeling of waste and repository characteristics to show that they could meet their disposal requirements. Ensuring that the Idaho calcine can be disposed without further treatment would provide a strong cost driver for this R&D. Processing Idaho calcine to produce a different waste form would likely require a DOE investment of several billion dollars. • Work that is synergistic with WP-4. New melter technology might, for example, allow waste processing at higher temperatures, which could increase waste loading and provide faster throughput as well. 

44 ADVICE ON THE DOE’S CLEANUP TECHNOLOGY ROADMAP Prioritization of the Gap Relative to other science and technology gaps discussed in this section the committee prioritized this gap as Medium. Relative Rating Criteria High Medium Low Volume of waste affected X Potential to reduce technical uncertainty X Potential to affect cleanup schedule X Potential to affect cost X WP-6: A variety of wastes and nuclear materials do not yet have a disposition path. Wastes and nuclear materials that do not have a defined disposition path are “orphans.” Examples of orphans include: • Over 4,000 cubic meters of calcine at Idaho that may require pro- cessing such as vitrification if it cannot be shown to be acceptable for in situ disposal or geologic disposal in its present form; • Idaho’s sodium-bearing tank waste, which is presently classified as TRU waste;10 • The waste left in SRS Tank 48 that contains tetraphenyl borate (used in a previous attempt to remove Cs-137 from tank waste) and its degradation products, which may require special processing to convert them into a stream suitable for vitrification at the DWPF; • Spent fuel at Idaho for which adequate characterization to qualify it for disposal is impractical because of the high radiation field and lack of access to the nuclear materials in the sealed packages; • Aluminum-clad N-reactor fuels at Hanford that may not meet cri- teria for disposal in a deep geologic repository because of their susceptibil- ity to corrosion; • K-basin sludge at Hanford that contains pyrophoric uranium metal; • High-atomic-weight (“heavy”) actinide targets at ORNL that have no further use; • The used beryllium neutron reflectors at the High Flux Irradiation Reactor at ORNL and Advanced Test Reactor at INL, which are classified as civilian rather than as defense waste. Their content of TRU radionuclides from uranium impurities and/or carbon-14 from nitrogen impurities puts 10 Idaho currently has a project for treating these wastes, but they will have to be stored onsite until their final disposition is decided. 

PRINCIPAL SCIENCE AND TECHNOLOGY GAPS 45 them in the civilian greater-than-Class C (GTCC) waste category. Wastes in this category presently have no disposition path (NRC 2006a). These wastes and nuclear materials have significant radioactivity and therefore hazards associated with them. Some of the wastes constitute a substantial volume or number of items that must be dispositioned. There may be enough variability in composition or other properties among wastes in any of the categories above to make their characterization a challenge. Impact of Gap The existence of this gap has two ongoing and potential impacts: • Continuing cost and occupational doses result from having to con- tinue to operate the storage facilities and associated site infrastructure. • Until stabilized, many of these materials will continue to degrade, especially those stored underwater, which could increase the cost and haz- ard of retrieval, treatment, and disposition. Current Status Alternatives for disposition of the INL calcine and Tank 48 waste are being evaluated but the unique nature of these wastes and limited previous R&D suggest that a fundamental understanding of these materials is not in hand. R&D is under way to prevent future beryllium reflectors from becoming GTCC waste. Approaches to Bridge the Gap The NRC has previously recommended R&D (NRC 1999a, 2001b, 2003, 2005, 2006b), but little has been done to determine the final disposal route for these wastes. According to this committee’s assessment of infor- mation it received, the following have promise for future EM R&D: • A systematic effort to develop the technical basis for alternative characterization, treatment, and disposal options, and for waste acceptance criteria; • A systematic effort to understand the degradation rate of nuclear materials in storage with initial focus on materials stored underwater; • Risk-informed comparison of the alternatives for disposition of INL calcine and SRS Tank 48 waste; and • Improved methods for characterizing highly radioactive spent fuel and nuclear materials inside containers. 

46 ADVICE ON THE DOE’S CLEANUP TECHNOLOGY ROADMAP Prioritization of the Gap Relative to other science and technology gaps discussed in this section the committee prioritized this gap as Low. Relative Rating Criteria High Medium Low Volume of waste affected Xa Potential to reduce technical uncertainty X Potential to affect cleanup schedule X Potential to affect cost X aIfIdaho calcine in its present form is determined to be unacceptable for disposal. GROUNDWATER AND SOIL REMEDIATION The groundwater and soil remediation program area of the EM road- map deals primarily with environmental sampling and contaminant char- acterization, treatment, remediation, and modeling to guide cleanup. EM is responsible for about 6.4 billion cubic meters of contaminated soil, groundwater, and other media that may require remedial action (NRC 2000c, p. 24). Chemicals, metals, and radionuclides were introduced into the environ- ment at DOE sites through accidental spills and leaks from storage tanks and waste transfer lines and also through intentional disposal via injection wells, disposal pits, and settling ponds. Releases into the environment gen- erally were not closely tracked, and many release sites were unmarked and forgotten. Some of these sites are being rediscovered as EM proceeds with its cleanup program. Chlorinated hydrocarbons or volatile chlorinated organic compounds are the most prevalent group of contaminants at DOE sites, appearing with a frequency of 82 percent in plumes reported in the EM Ground Water Database (GWD).11 These compounds were used in large quantities as cleaning agents, solvents, or lubricants. Some of these compounds, which are sparingly miscible with water and denser than water—notably carbon tetrachloride—comprise a category of contaminants referred to as dense, nonaqueous phase liquids (DNAPLs). Organic contaminants are often co- contaminated with tritium or nitrates (Hazen et al. 2008). Plutonium, uranium, strontium, technetium, chromium, and mercury are among the problematic radionuclides and toxic metals that have been difficult to predict in occurrence and transport. While metals and radionu- 11 See http://www.em.doe.gov/Pages/groundwatersoildatabase.aspx?PAGEID=DB. 

PRINCIPAL SCIENCE AND TECHNOLOGY GAPS 47 clides (other than tritium) are reported to occur in only about 5 percent of the plumes in the GWD, they occur together nearly 30 percent of the time and are associated with volatile organic chemicals (VOCs) approximately 25 percent of the time (Hazen et al. 2008). Different contaminants com- bined into clusters of three or four are not uncommon, and they frequently include radioisotopes, metals, sulfates, and nitrates. Such combinations probably reflect their origins from onsite chemical operations and subse- quent interactions within the geologic media. In 2004, a DOE Inspector General’s audit (DOE 2004c) found the con- tinued use of pump-and-treat technology to be relatively ineffective, that in- novative groundwater contaminant monitoring is not being exploited, and that implementing current treatment and barrier technology may need abey- ance until realistic end states are more sharply defined. Thompson (2007) reported that a 2006 audit by the Government Accountability Office found fault with DOE’s remediation efforts to prevent contaminants from reach- ing the Columbia River. The audit (GAO 2006) concluded that technology used in several remedies is not performing satisfactorily, and that there is a lack of new technologies to address contamination issues. Groundwater remediation challenges are recognized in previous NRC (1993, 2000b, 2004) reports and were reiterated in the committee’s workshop summary (NRC 2007c). GS-1: The behavior of contaminants in the subsurface is poorly understood. Geochemical and biochemical oxidation and reduction of metals and radionuclides can dramatically alter their solubility in groundwater, sorption on solid substrates, and colloidal transport properties. These are complex, dynamic, and often reversible processes that cannot be predicted without knowledge of the contaminant chemistry, the subsurface biogeochemical and hydrogeologic properties and their spatial variability, and dynamics of water recharge and removal (e.g., precipitation, stream flow, groundwater pumping). Lack of basic understanding of contaminant and site characteristics can lead to incorrect concepts of contaminant behavior that have, in turn, led to a disconnect between the expected and actual outcome of remediation efforts. Several examples illustrate the importance of understanding basic contaminant biogeochemistry and characterizing the properties of the field site adequately when planning whether or how to conduct soil and ground- water remediation: • At Oak Ridge, the levels of mercury in East Fork Poplar Creek, which is downstream from Y-12, have been reduced to meet drinking wa- 

48 ADVICE ON THE DOE’S CLEANUP TECHNOLOGY ROADMAP ter standards by cleanup actions already completed. However, the mercury concentration in fish and aquatic life is continuing to increase (Appendix F), suggesting a critical aspect of the contaminant distribution, transport, and/or biotransformation at the field site is not known. • Initial predictions of contaminant migration to the water table at Idaho’s RWMC developed in the 1960s were on the order of 100,000 years. Improved knowledge of subsurface transport processes has led to travel time estimates that are on the order of decades (NRC 2000c, p. 30). • Stewart (2007) reported seven examples of apparently anomalous contaminant migration at Hanford—the contamination was moving in unexpected amounts and/or unexpected directions. The reasons underlying the apparently anomalous behavior were resolved in each case by scientific study that led to improved approaches for remediation or containment of the contamination. However, she also noted problems for which scientific understanding is limited. One example is the deeper migration of plutonium into the vadose zone beneath the Z cribs than has been predicted with cur- rent site models. • In the Hanford 300 area, uranium-contaminated soil was removed, and the plume was expected to meet the water quality standard within 10 years of the remediation. This did not happen. Incomplete characteriza- tion of the source zone and the consequent lack of a remedy to deal appro- priately with the source zone have prevented EM from meeting the original cleanup time line. In addition, the assumption that the groundwater plume would dissipate during the 10 years following soil removal delayed further progress to understand the source of what is now known to be an ongoing groundwater contamination issue. • The unexpected detection of chromium in a monitoring well at Los Alamos substantially reinforced concerns about the adequacy of that site’s groundwater protection program (NRC 2007b). Impact of the Gap If the rate of progress or result of a remedial action is less than ex- pected, it can delay schedules (including missed regulatory milestones), increase costs, and undermine stakeholder confidence in EM’s site cleanup. Remediation carried out without complete (or ongoing) characterization of the contamination source and factors that control contaminant movements is a technical risk—simply removing some of the contaminant mass may not cause the expected response in contaminant concentration or movement. Without knowledge of the fundamental processes that interact to determine contaminant mobility and persistence in the spatially heterogeneous geo- logic settings that exist at each of the DOE sites, it is impossible to complete a reliable risk assessment or plan an effective remedial program. 

PRINCIPAL SCIENCE AND TECHNOLOGY GAPS 49 Current Status There is substantial R&D under way at the DOE sites and by many other organizations to better understand the science needed to predict how various hazardous substances are mobilized, transported through the geosphere and biosphere, and affect humans and the environment. For ex- ample, more detailed characterization of the source of the uranium plume in the Hanford 300 area is included in one of three Integrated Field-Scale Subsurface Research Challenges funded through the Environmental Reme- diation Science Program (ERSP) within the Basic Energy Sciences Office of DOE’s Office of Science (SC).12 Work to improve conceptual and computational models of contaminant migration is ongoing within the Subsurface Science Focus Area through sup- port by the Environmental Remediation Sciences Division of the DOE Office of Science’s Office of Biological and Environmental Research (OBER).13 Examples of Subsurface Science Focus Area projects include PNNL’s re- search on the role of microenvironments and transition zones in the reactive transport of technetium (Tc), uranium, and plutonium;14 ORNL’s research on the biogeochemical transformations that govern mercury speciation at the sediment–water interface;15 and LBNL’s research to develop a sustainable systems approach for addressing critical knowledge gaps associated with envi­ ronmental stewardship of metals and radionuclides in the subsurface.16 Organizations such as the DOE offices of Science and Civilian Radioac- tive Waste Management, the Environmental Protection Agency (EPA), the Nuclear Regulatory Commission (USNRC), and the U.S. Geological Survey all have technical expertise relevant to assessments and could be helpful in establishing generic approaches and assumptions. The EPA and USNRC both do technical assessments to support decisions that include substantial stakeholder involvement, and their experience could be useful to the DOE cleanup program. The EPA Superfund program established to support cleanup and risk reduction of those sites on the National Priorities List (NPL) and the Su- perfund Innovative Technology Evaluation (SITE) program are helpful re- sources in addressing what works and what does not for problematic sites and in deployment of new technology. While some contaminants and discharge magnitudes are unusual or unique to DOE sites, many industries and universities have major R&D programs for understanding contaminant fate and transport. 12 See http://www.hanfordifc.pnl.gov. 13 See http://ersdprojects.science.doe.gov/. 14 See http://www.pnl.gov/biology/sfa/. 15 See http://www.esd.ornl.gov/programs/rsfa/index.shtml. 16 See http://esd.lbl.gov/research/projects/sustainable_systems/. 

50 ADVICE ON THE DOE’S CLEANUP TECHNOLOGY ROADMAP Approaches to Bridge the Gap According to the committee’s assessment of information it received, the following have promise for future EM R&D: • Development of improved technology and methodologies for source and plume characterization and monitoring with emphasis on real-time analytical monitoring instruments for field use, remote data acquisition, and automated data collection; • Continued study of contaminated materials from vadose zone and groundwater plumes to improve conceptual models of geochemical and biogeochemical processes (and species) controlling the mobilization and immobilization of contaminants, including complex chemical mixtures; • Use of more sophisticated computational models that better incor- porate understanding of site geohydrology and contaminant geochemistry; and • Development of scientific bases to support delaying remediation activities until there is an adequate knowledge base to proceed with the remediation. The technical challenges in groundwater and soil remediation differ from those in waste processing in terms of the timescales during which the relevant processes operate, access and ability to measure process parameters (reading a gauge versus ascertaining what is going on belowground), and ability to control the process parameters. The need for adequate charac- terization of site hydrogeology and contamination sources and plumes is recognized by EM and SC. Partnering with SC (Chapter 4) can provide EM with access to state-of-the-art science capabilities to improve its site charac- terization through more robust modeling and advanced instrumentation. In the case of the EPA Superfund program, each NPL site has a Record of Decision (ROD) that describes how the given site will be cleaned up or managed. The RODs are readily accessible public documents that may translate to EM site closure. Perhaps more informative are the Remedial Investigation/Feasibility Study reports that lay out the investigative assess- ment of a site and the treatability studies required to select and support site cleanup. The congressional language that led to the EM roadmap cited a pre- vious NRC (2005) report, which found that “monitoring systems at EM closure sites have been estimated to be some 25 years behind the state- of-art.” and that “an improved capability for environmental monitoring would strengthen EM's plans to leave waste and contaminated media at DOE sites.” Two NRC reports (2005, 2007b) suggested R&D in nonin- vasive geophysical sensor techniques such as electromagnetic and electrical 

PRINCIPAL SCIENCE AND TECHNOLOGY GAPS 51 resistivity methods, seismic reflectivity, and ground-penetrating radar to r ­ educe, as much as possible, the practice of physically sampling and analyz- ing groundwater samples, which is currently prevalent at DOE sites. It has been argued that the complexity of the sites, microlevel of resolu- tion needed, and integration among models require computational capacity that may not be justifiable or that model complexity is too great to provide much real value to practitioners. While this may have been true in the past, advances in both the computational capability within DOE and risk assess- ment modeling clearly indicate that new modeling efforts in mixed contami- nant fate and transport analysis, in the development of methods for scaling up from micromeasurements to field-scale prediction, and in the simulation of remediation processes are promising. This computational power could be linked to the development of improved methods for characterizing site conditions, formulating conceptual models that represent system behavior, parameterization and calibration of site-specific models, and quantification of uncertainties in prediction. Modern computing power can help ensure that more sophisticated numerical models are well integrated with the biochemical, ecological, and geochemical sciences sufficiently to provide the resolution needed to improve the accuracy of model simulations and predictions needed to advance cleanup, remediation, and risk reduction. For example, promising work at the Environmental Molecular Sciences Laboratory at PNNL in the use of instruments such as nuclear magnetic resonance to generate microscale data on intragrain diffusion rates of dis- solved uranium in tank-waste-contaminated sediment particles suggests that new modeling efforts can be appropriately parameterized. There are some instances of which today’s understanding of site and contaminant characteristics and/or available technologies are probably not adequate for a successful remediation program. The BC cribs on the Central Plateau at Hanford are likley examples. The liquid waste disposed at the BC cribs and trenches represents some of the most concentrated radioactive and hazardous waste disposed to the ground at Hanford. Based on inven- tory estimates, this site contains the largest inventory of technetium-99 in the Hanford soil.17 The majority of the Tc-99 is believed to be located in the site’s vadose zone, which comprises highly stratified glacial-fluvial sedi- ments that give rise to complex subsurface-flow paths (Gee et al. 2007). This complexity, combined with uncertainty about the in-ground chemical and/or biological processes that influence Tc-99 behavior, means that the crucial information needed to design appropriate remedial actions for the site is missing. Another example is the chlorinated organic contamination at the East Tennessee Technology Park at Oak Ridge (Appendix F). This case involves 17 See http://www.hanford.gov/cp/gpp/functionalareas/wastesite/bccribs.cfm. 

52 ADVICE ON THE DOE’S CLEANUP TECHNOLOGY ROADMAP the presence of a DNAPL in a fractured bedrock aquifer, which is a com- bination of contaminant and subsurface characteristics that is universally acknowledged to be extremely challenging to effectively remediate. Oak Ridge listed a need for scientific and technical support for a Technical Impracticability (TI) waiver for remediating these source areas (Phillips 2007). A decision to grant a TI waiver represents regulators’ concurrence with a finding that restoration of contaminated soil and/or groundwater to agreed cleanup levels cannot be achieved using currently available or new and innovative methods or technologies.18 For such cases in which current knowledge and technology are likely insufficient to ensure successful remediation, an alternative approach is the “cocooning” concept, which is being used for the Hanford reactors (NRC 2005). The concept of cocooning for soil and groundwater contamination problems that cannot be technologically addressed at present is to adap- tively manage the contamination in order to avoid actions that involve costly or inappropriate treatment activities and result in little to no risk reduction. Developing the science and technology base to show that a temporiz- ing measure, such as pumping strategies to achieve hydraulic containment that prevents the spread of a plume (GS-2) or placing a water barrier over a trench (GS-3), is protective while continuing to pursue better solutions as the state of the art advances could help EM deal with currently intractable situations. R&D to demonstrate that the current situation is safe, for now, and to develop a temporizing remedy could save money and time and avoid the perception of failure. Roadmapping the longer-term research to eventu- ally address the problem could help assure stakeholders that the problem is not simply being pushed aside. Prioritization of the Gap Relative to other science and technology gaps discussed in this section the committee judged the priority of addressing this gap as High. Relative Rating Criteria High Medium Low Volume of waste affected X Potential to reduce technical uncertainty X Potential to affect cleanup schedule X Potential to affect cost X 18 See: http://homer.ornl.gov/oepa/guidance/cercla/techimpract.pdf. 

PRINCIPAL SCIENCE AND TECHNOLOGY GAPS 53 GS-2 Site and contaminant source characteristics may limit the usefulness of EM’s baseline subsurface remediation technologies. There are a wide variety of methods for remediating contaminated groundwaters and soils that have been developed and used by private in- dustries and government agencies. These include technologies such as pump and treat, biostimulation and bioaugmentation, air-sparging, soil vapor extraction, electrokinetics, phytoremediation, in situ flushing and in situ oxidation, permeable reactive barriers (PRBs), in situ thermal treatment, multiple-phase extraction, and monitored natural attenuation (MNA). Among these, pump and treat is the most commonly used, and it appears to be favored by EM site cleanup contractors (Figure 2.6). It is a mature technology, frequently used for remediating groundwater contaminated with a variety of substances, including VOCs, residues of explosives, and dissolved metals. Contaminated groundwater is removed from the subsur- face by pumping, treated to remove the contaminants, and returned to the aquifer or discharged. The water well design, pumping system, and treatment depend on the site characteristics and contaminant type. Aboveground treatment technolo- gies for extracted contaminated groundwater typically include biodegrada- tion, filtration, air stripping, and adsorption. It is not uncommon to find many wells extracting groundwater at the same time. These wells may extract water from different depths to maximize effectiveness. Groundwa- FIGURE 2.6  Illustration of groundwater remediation system at Hanford. In a Fig 2-6 typical pump-and-treat system, like that featured in this illustration, water wells bitmap image intercept a contaminant plume and pump water to the surface where it is treated to remove the contamination. Treated groundwater is usually returned to the aquifer. SOURCE: Department of Energy. 

54 ADVICE ON THE DOE’S CLEANUP TECHNOLOGY ROADMAP ter wells also provide a way to monitor progress of the remediation and to make adjustments to the system in response to changes in subsurface conditions. In many circumstances pump-and-treat systems can be effective in preventing further migration of plumes, but they may be ineffective for eliminating them (Sidebar 2.2). Because aquifers do not have uniformly permeable strata (or uniform biogeochemical properties) pumping systems cannot uniformly remove contaminants. For example, in a sedimentary aquifer system comprising fine-grained silts interbedded with sands, most of the flow occurs in continuous sandy zones and little flow moves through the silts due to the permeability difference. Contaminants are flushed relatively quickly from the high-permeability zones while contamination in the low- permeability zones can remain largely in place. Once the high-permeability zones are flushed, the concentration gradient drives contaminants to diffuse from the low-permeability (high concentration) to the higher-permeability zones, where their concentration is diluted by greater flux. SIDEBAR 2.2 Carbon Tetrachloride Plume Remediation at Hanford As the preferred remedy identified in the draft feasibility study for the carbon tetrachloride (CCl4) plume at the 200 West Area, Hanford recommended a pump-and- treat (P&T) system, already partially in place as part of an interim remedy, combined with flow path controls, institutional controls, and monitored natural attenuation. The purpose of the interim remedy for the ~11 km2 CCl4 plume (estimated area exceeding the water quality standard in the upper region of the unconfined aquifer) is to prevent further migration of the portion exceeding 2,000 μg/L. To date, the P&T remedy appears to be successful in achieving the stated goal. From 1994 to 2006, the pump and treat system has removed approximately 10,000 kg of CCl4 through the removal and treatment of over 3 billion liters of water using up to 10 extraction and 3-5 injection wells. For reference, the estimated mass of CCl4 discharged to ground, mostly as a DNAPL oil mixture, was 577,000 to 922,000 kg. Additional extraction wells are planned. Maps of the CCl4 plume comparing the extent in 2006 to 1990 show that the 1,000-μg/L and 2,000-μg/L contours for the upper portion of the aquifer encompass smaller areas in 2006 compared with 1990. However, the portion of the plume at lower concentrations (<1,000 μg/L) has continued to expand. Initial characterization of plume extent focused only on the surficial portion of the aquifer. Recent characterization has shown that CCl4 is present at concentrations above the water quality standard throughout the thickness of the unconfined aquifer (~60 meters thick). High concentrations are found at depth and also displaced from the high concentrations at the surface. The impact of P&T to date on the deeper portion of the plume cannot be evaluated. The recent three-dimensional characterization has 

PRINCIPAL SCIENCE AND TECHNOLOGY GAPS 55 Pump-and-treat systems in heterogeneous aquifers (as are most) typi- cally produce a two-stage mass extraction behavior over time that has been attributed to permeability heterogeneity in the subsurface, as described above, in at least some cases. Rapid flushing of more permeable zones and high initial mass extraction rates are followed by tailing behavior typified by sustained contaminant concentrations (and mass removal) that change slowly. Differences in sorptive capacity among units or among grains of the geologic media can lead in a parallel way to variability in the capacity to store contaminants in the solid phase and similar types of behavior. The high concentration regions that remain within the groundwater system can cause the contaminant concentration to rebound when a pumping system is turned off. Concentration rebound following the cessation of pumping can also result from increased access to contaminant mass stored above the pumped water table in an unconfined aquifer. resulted in a significant increase in the estimated CCl4 mass present in the unconfined aquifer. The initial concentrations of CCl4 present in the groundwater and the detections of CCl4 throughout the aquifer thickness are both consistent with transport of DNAPL- containing CCl4 oil below the water table. Experience with this project points to scientific and technical challenges that are inherent in groundwater and soil remediation. The groundwater flow path to the Colum- bia River is very long, allowing for significant reaction between CCl4 and the aquifer solids. In order to constrain risk predictions from a potential continuing source of CCl4, characterization of the aquifer and groundwater biogeochemical and hydrogeologic sys- tem, as well as reactions with CCl4, is needed. The uncertainty surrounding all aspects of the CCl4 plume (presence of DNAPL, aquifer reactivity) was previously unknown, leading to extreme variability in risk assessment outcomes that drive the cleanup. Some recent activities surrounding the CCl4 plume provide a positive model for reducing uncertainty, including improved characterization of the plume, for example. In addition, some research into CCl4 reactions with aquifer solids, previously not evaluated, has recently been provided limited support by ERSP. However, the aquifer itself is relatively poorly characterized, especially with respect to the biogeochemical conditions that can impact the fate of CCl4 and many other mobile groundwater contaminants relevant to EM at the field site, such as nitrate, uranium, and technetium. Similar to some other DOE sites in the western United States, the aquifer is deep and the geologic system is challenging to sample. Currently available methods do not provide robust information. There are new ERSP projects that seek to provide improved geologic characterization of contaminated sites through the use of geophysical tools. Improvement in methods of site characterization and monitoring in deep and/or challenging geologic systems offers the potential for significant benefit at DOE sites. 

56 ADVICE ON THE DOE’S CLEANUP TECHNOLOGY ROADMAP Impact of the Gap Because of the heterogeneous distribution of physical and biogeochemi- cal properties in groundwater, groundwater is inherently a poorly mixed system over short travel distances. As a consequence, pump-and-treat sys- tems and other active remediation methods (methods that involve actively processing the groundwater in order to remove the contaminants) may be inefficient or ultimately ineffective. Operating these processes over the long time periods that may be required (e.g., 50-100 years) to meet cleanup goals can be expensive, and they may be discontinued or considered to have failed before the goals have been met. Premature implementation of pump and treat as the baseline technology can divert resources away from finding less expensive and ultimately more effective solutions. Current Status At Oak Ridge, DOE has used a continuous pump-and-treat system at the east end of Y-12 to keep an underground plume of carbon tetrachloride from spreading farther. Water is pumped to the surface, treated, and then released into a nearby creek. This is a large plume that is evidently being fed from an underground source of the carbon tetrachloride. Although it has been effective in limiting the plume’s offsite migration, the treatment system has not eliminated the source or significantly reduced the concentration of carbon tetrachloride in the plume (Appendix F). A similar situation exists at Hanford, where pump and treat has controlled a carbon tetrachloride plume but impact on the source and plume longevity is unknown (Sidebar 2.2). Whitaker (2008) described successful applications of two active reme- diation approaches at SRS that may be improvements over pump and treat. One is a steam injection and contaminant removal system that is remediat- ing a 3-acre area regarded as the primary source of subsurface contamina- tion in A- and M-Areas. This dynamic underground stripping system is expected to complete the remediation in 5 years versus an estimated 200+ years using conventional technologies. The system reportedly had removed 380,000 pounds of solvents at the time of the committee’s visit. Second was the use of electrical resistance heating, which removed 710 pounds of solvents at C-Reactor in 2006. The system achieved 99 percent efficiency according to soil samples, and completed the cleanup 2 years faster than soil vapor extraction. In 2007, the aboveground equipment was relocated to an area referred to as the CMP pit where chemicals, metals, and pesti- cides were disposed (Whitaker 2008) Although active remediation measures such as these at SRS can be effective, several NRC reports (1994, 1997b, 2000b) point to potential 

PRINCIPAL SCIENCE AND TECHNOLOGY GAPS 57 limitations of active remediation approaches. Three reports (NRC 1994, 1997b, 1999b) recommended that additional work be undertaken on pas- sive remediation technologies. Passive barriers limit contaminant flux by reducing concentrations through biological or chemical reactions. As for any barrier system, barrier longevity is a critical aspect of success. The In Situ Redox Manipulation (ISRM) barrier that was installed to remediate a chromium groundwater plume in Hanford’s 100­D Area is one case in point. Laboratory experiments performed before installation of the ISRM barrier indicated that it would be effective for approximately 20 years, but localized signs of failure were discovered after only 18 months. The cause of premature barrier breakdown was determined to be heteroge- neities in the aquifer, where laterally discontinuous units with high permea- bility and lower inherent reductive capacity (because of lower iron content) were reoxidized faster than the less transmissive layers (DOE 2004a,b). In the mid 1990s, OBER formed the Natural and Accelerated Biore- mediation Research program to develop more fundamental insight into the interplay between geochemical and biological processes that may lead to ef- fective and new bioremediation technology. This program consumed much of the OBER Subsurface Science Program and in 2005 was merged with the Environmental Management Science Program to create the ERSP, which supports fundamental, mission-oriented research on DOE legacy waste and priority contaminants. These research programs have generated nearly 1,000 research publications relative to the mechanistic microbiology, fate, and transport issues influencing metals and radionuclides in the subsurface and their potential for bioremediation and immobilization. Approaches to Bridge the Gap According to the committee’s assessment of information it received, the following remediation approaches have promise for future EM R&D: • Reactive chemical barriers, • Natural attenuation, and • Bioremediation. Approaches such as natural attenuation (Sink et al. 2004) and PRBs and treatment zones19 may provide EM with remediation solutions that are lower cost and more likely to succeed over the long term than baseline approaches such as pump and treat. To enhance its use of MNA, EM needs a better technical basis for determining the situations for which it is an appropriate tool for reme- 19 See http://clu-in.org/download/rtdf/2-prbperformance_web.pdf. 

58 ADVICE ON THE DOE’S CLEANUP TECHNOLOGY ROADMAP diation. This requires detailed understanding of contaminant associations, biogeochemistry, hydrology, and projections of future behavior. A recent EPA (2008) paper discusses site characterization to support use of MNA for remediation of inorganic contaminants in groundwater. In the area of PRBs and treatment zones, research is needed on pre- dicting and improving the lifetime of iron PRBs and the development and performance modeling/monitoring of non-iron-based systems. Several prior NRC reports provide information and recommendations associated with MNA and PRBs (NRC 1994, 1997b, 1999b, 2000b, 2007c, p. 15). The concept of cocooning subsurface contamination described in GS-1 involves stabilizing contamination in place for now; monitoring it until ra- dioactive decay, other natural processes, or new technologies make ultimate cleanup feasible or unnecessary; adapting to new knowledge, which may accumulate from R&D over years or decades; and making DOE’s long-term responsibilities clear to all stakeholders. Bioremediation approaches may provide cost-effective remediation op- tions under conditions where other options are not feasible (e.g., aerobic, fractured bedrock at Oak Ridge, “tight” formations at SRS, and deep vadose zones at Hanford and Idaho). The challenge for bioremediation, however, is that most of the contaminants are either degraded slowly and/or incompletely, and they require use of growth-supporting substrates for gen- eral microorganisms. The organisms that efficiently degrade contaminants such as trichloroethylene, the predominant VOC at DOE sites, are very specific. Depending on whether the environment is aerobic or anaerobic, completely different organisms and biochemistries operate. While many metals, minerals, and radionuclides may be biochemically oxidized or reduced by microorganisms, this capacity is again very specific to select groups of individuals or species. At Hanford, tests are under way to introduce lactate to promote subsurface anaerobic, hexavalent chromium reduction and immobilization. That work is being coupled to state-of-the- art gene expression monitoring to prove the in situ physiological basis of the process. While new organisms that are capable of reducing contaminant con- centrations have been found through DOE-supported research, they may not be predominant in the environment. Controlling their biochemical activity for actual application in groundwater remediation would be dif- ficult without adequate fundamental understanding of the biogeochemical conditions of the specific area to be remediated and the ability to monitor the processes involved. Given the breadth of site conditions and contami- nants of concern that DOE manages, key targets for remediation would be emphasized, such as the carbon tetrachloride plumes at Hanford and INL. These area-specific characterizations would be coupled to the development of methods for obtaining in situ biogeochemical information, fine-scale 

PRINCIPAL SCIENCE AND TECHNOLOGY GAPS 59 geophysical characterization, and high-resolution in situ monitoring. Subse- quently, these investigations can improve the conceptual site models that are used to integrate the fine-scale structure, transport, and chemical reactivity that are needed to guide transport predictions and process optimization for site remediation. Prioritization of the Gap Relative to other science and technology gaps discussed in this section the committee judged the priority of addressing this gap as Medium. Relative Rating Criteria High Medium Low Volume of waste affected X Potential to reduce technical uncertainty X Potential to affect cleanup schedule X Potential to affect cost X GS-3: The long-term performance of trench caps, liners, and reactive barriers cannot be assessed with current knowledge. Engineered containment barriers, such as trench caps, liners, and reac- tive barriers, are designed to reduce risks associated with buried wastes and subsurface contamination by preventing the spread of contamination and/or minimizing the amount of surface- and/or groundwater that comes into contact with the wastes and contamination (Figure 2.7). On the basis of as many as 20 years of observations, a recent NRC (2007a, p. 1) report that assessed the performance of engineered waste containment barriers concluded that “most engineered waste contain- ment barrier systems that have been designed, constructed, operated, and maintained in accordance with current statutory regulations and require- ments have thus far provided environmental protection at or above speci- fied levels.” The report (p. 2) also stated that although extrapolations of long-term performance can be made from existing data and models, such e ­ xtrapolations will have “high uncertainties until field data are accumulated for longer ­periods, perhaps 100 years or more.” In the case of natural ana- logue systems, such as the Fernald mixed-waste landfill in Ohio, which are designed to serve 1,000 years or more without regular maintenance, the report (p. 73) concludes that “maintenance-free covers have not yet been demonstrated to work.” Thus, the long-term performance of containment barriers, such as trench caps, liners, and lateral walls, cannot be assessed given the current state of knowledge. 

60 ADVICE ON THE DOE’S CLEANUP TECHNOLOGY ROADMAP FIGURE 2.7  Diagram of an engineered, near-surface, waste disposal facility. Similar approaches, all of which depend on special materials and construction designs, are used for permanent disposal of low-level radioactive waste and hazardous materials Fig 2-7 at the DOE sites the committee visited. Such engineered facilities are intended to contain the waste for hundreds of bitmapmore. years or image SOURCE: Department of Energy. Impact of the Gap Removal and subsequent treatment of wastes at many DOE sites is technically difficult, expensive, and potentially hazardous to workers. As a result, alternative approaches that leave the waste in place, but incorporate robust containment barriers and waste stabilization technologies over the long term (100-1,000 years), are a key element of DOE’s strategy for man- aging legacy waste sites. Without confidence in predicting the performance of containment barriers beyond a few decades, many of DOE’s performance assessments, which assume long-term barrier integrity, could be deemed unreliable enough to prevent current plans for area and site closures go- ing forward. Repair or replacement of engineered barrier systems that fail in a relatively short time could increase costs, delay site closure, and raise stakeholder concerns about the likelihood of future failures. Current Status The behavior of engineered barrier systems is influenced by environ- mental and ecological conditions that will continue to evolve over time as a result of processes like ecological and biological successions, landform evolution, and climate change. Research in the INL Environmental Sur- veillance, Education and Research Program20 and at the Savannah River 20 See: http://www.stoller-eser.com/research.htm. 

PRINCIPAL SCIENCE AND TECHNOLOGY GAPS 61 Ecology Laboratory (SREL)21 is aimed at understanding such processes, thus providing information relevant to the robust design, construction, and maintenance of containment barriers. A Hanford-designed prototype surface barrier, referred to as the Han- ford barrier, is a 2.5-hectare multilayered, vegetated, capillary barrier com- posed mainly of stable natural materials and designed to isolate buried wastes for about 1,000 years. While not all near-surface disposals at Han- ford will require the degree of protection offered by the Hanford Barrier, Ward (2007) stated that the results of tests and monitoring of the barrier’s performance can be used to guide the design of more modest covers. To do so, it will be necessary to determine moisture flux through representative waste sites, including vegetated and graveled surfaces, account for seasonal variations in precipitation and heating, and from this information develop robust infiltration barriers for sites where contaminants will be left in place (Appendix D). The Alternative Cover Assessment Program (ACAP) is developing field- scale performance data for landfill final cover systems based on field data being obtained at a dozen sites representing a variety of geohydrologic con- ditions.22 ACAP is part of the EPA’s National Risk Management Research Laboratory’s SITE program established to promote the development of new and innovative technologies used to address hazardous waste problems. Both prescriptive (Resource Conservation and Recovery Act) and innova- tive alternative cover designs are currently being tested in the project. SREL is investigating the alternative use of native grasses for vegetated caps (Kwit and Collins 2008), while projects at INL are exploring evapo- transpiration cap designs (ET barriers) as a low-cost, low-maintenance alternative to traditional designs.23 EM’s Office of Engineering and Tech- nology recently hosted a workshop on landfills (Benson et al. 2008). Approaches to Bridge the Gap According to this committee’s assessment of information it received, the following approaches have promise for future EM R&D: 1. Monitoring systems that can provide information on containment barrier performance can (i) reduce uncertainty related to the long-term performance of such engineered controls and (ii) lay the foundation for approaches that can provide early warning of unexpected or unacceptable barrier behavior so repairs or adjustments can be made in a timely fash- 21 See: http://www.uga.edu/~srel/. 22 See http://www.acap.dri.edu/. 23 See: http://www.stoller-eser.com/NERP/PCBE.htm. 

62 ADVICE ON THE DOE’S CLEANUP TECHNOLOGY ROADMAP ion. Monitoring systems might include buried sensors, surface or airborne surveillance, eco- and/or bio-indicators, and software support. Robust and low-cost systems that reduce manual labor and allow for remote, real-time access to data on barrier performance offer the most promise for improved monitoring. 2. Robust models of barrier behavior that can incorporate appropri- ate uncertainty and account for natural and anthropogenic spatial and temporal changes, together with field data to calibrate these models, can better assess long-term barrier behavior. To identify unacceptable barrier behavior, a scientific basis for what is an unacceptable “barrier breach” would be needed. 3. Many of the barrier systems put in place or proposed at DOE sites are systems that are designed to shed precipitation and/or divert or retard groundwater flow. Thus, they are systems that are intended to resist natural processes rather than work with them. These systems cannot be expected to provide long-term waste or contaminant isolation without continued maintenance or, in some cases, replacement and remediation at considerable effort and cost. The continued development and performance monitoring of alternative systems, such as natural analogues to existing landscapes and ET barriers that can work with rather than against nature, are thus needed (Clarke et al. 2004). 4. Engineered barriers, including those described in approach 3, above, might be better recognized as temporizing measures to control contaminant spread for years or decades until new technologies or natural processes provide a final solution. According to the cocooning concept introduced in GS-1, R&D would first be directed at ensuring safety of the engineered barrier system, with continued R&D toward improved charac- terization, monitoring, and modeling to ensure safe, permanent disposition of the contaminants or the contaminated area. 5. For the barrier designs themselves, R&D could be directed at developing a design strategy, and associated life-cycle cost model, that involves “perpetual periodic replacement” of the cover or barrier system rather than a design philosophy based on barriers that need to last as long as practicable. Prioritization of the Gap Relative to other science and technology gaps discussed in this section the committee judged the priority of addressing this gap as Medium. 

PRINCIPAL SCIENCE AND TECHNOLOGY GAPS 63 Relative Rating Criteria High Medium Low Volume of waste affected X Potential to reduce technical uncertainty X Potential to affect cleanup schedule X Potential to affect cost X GS-4: The long-term ability of cementitious materials to isolate wastes is not demonstrated. Very large volumes of cementitious materials will be used in the EM cleanup with the objective of protecting soil and groundwater by encapsu- lating wastes. Cementitious materials are among the world’s most widely used and best understood construction materials. In the EM cleanup pro- gram, their high-volume applications include: • Grouting of emptied HLW tanks and associated inter-tank transfer pipes; • Stabilizing LAW in large monoliths, such as the SRS saltstone, or in smaller containers (e.g., 55-gallon drums); and • Constructing disposal vaults or other structures. Cementitious materials are the best, and as a practical matter the only ones, available for these applications. However, ensuring that they can ef- fectively isolate waste for hundreds of years or more will be an ongoing scientific and technical challenge. Before the HLW tanks at Hanford, Idaho, and SRS can be closed, they are to be filled with a cementitious grout material, which has two purposes: (1) to encapsulate or otherwise reduce the mobility of the residual waste and (2) to stabilize the tank structurally to support the overburden. Han- ford has 177 tanks, Savannah River 51 tanks (2 of which have been filled with grout), and Idaho 11 tanks (8 of which have been filled with grout). These are large, typically 100,000- to million-gallon tanks, from which it is not possible to remove all of the contaminated material in the tanks; see WP-1. Stabilizing the residual waste in these tanks is essential. Currently there is almost no experience in cleaning and grouting inter- tank transfer pipes that were used to move waste among the waste tanks. Relative to their volume, some of the pipes, especially those that have be- come plugged, are likely to contain more residual wastes than the tanks and be considerably more difficult to fill thoroughly with grout. Salt waste in the tanks at SRS is to be processed and the resulting 

64 ADVICE ON THE DOE’S CLEANUP TECHNOLOGY ROADMAP LAW incorporated into a grout called saltstone; see WP-2. Depending on the specific constituents of the salt solution, the grout is formulated using appropriate proportions of portland cement, fly ash, ground granulated blast-furnace slag, water, and chemical admixtures. The grout is pumped into concrete vaults, where it hardens. Saltstone has a low oxidation-re- duction potential (Eh) to stabilize key radionuclides such as Tc-99 in less soluble forms to reduce the rate at which they would leach out or migrate in the groundwater (Rosenberger et al. 2005; Shuh et al. 2002). The salt- stone vault has a concrete roof and will eventually have an engineered cap over the entire installation (NRC 2006b). The vault is also a barrier to contaminant release. Impact of the Gap The successful grouting of wastes in tanks, pipes, and saltstone is as- sumed in performance assessments that demonstrate regulatory require- ments for tank closure and SRS salt disposal will be met. If the adequate long-term performance of the grout were to be seriously questioned—and the requirements for the grout’s performance and performance period are beyond any direct experience in the construction industry—then closure of the tanks and SRS salt disposal could become problematical. The Nuclear Regulatory Commission (USNRC 2005) Technical Evalu- ation Report (TER) on the DOE’s performance assessment of salt waste disposal at SRS (p. 50) indicates that the Commission has concerns about uncertainties in saltstone’s performance: In conducting its PA [performance assessment] of the facility, DOE con- sidered the various mechanisms of release to estimate the source term and release of contaminants. Both diffusive and advective transport processes were addressed. To model contaminant transport in the near field, there was a need to estimate the contaminant concentrations in the pore fluid based on the concentrations in the saltstone. However, relating inventory in the saltstone to pore fluid concentration is complicated by various processes, such as precipitation/dissolution reactions, aqueous complex formation, and sorption. DOE acknowledged that these processes are poorly understood and difficult to quantify for the SDF [saltstone disposal facility]. The TER (p. 52) summarizes the USNRC’s evaluation of the DOE’s model of saltstone and concrete vault degradation as follows: In general, the NRC staff agrees with the qualitative assessment of the deg- radation mechanisms for saltstone. However, given that: (1) the calculated releases from the SDF are sensitive to the values of hydraulic conductivity of the vault and saltstone; and (2) “the timing and extent of degradation 

PRINCIPAL SCIENCE AND TECHNOLOGY GAPS 65 are not readily predictable due to enormous uncertainties in conditions for thousands of years” (Cook and Fowler, 1992, Section 3.1.3.5), it would be useful to reduce the uncertainties associated with the hydraulic conductiv- ity and long-term integrity of the vault and saltstone. Additional labora- tory measurements of initial hydraulic conductivity, as well as long-term tests or monitoring studies designed to evaluate the long-term durability of the saltstone and concrete vault, would help reduce these uncertainties [italics added]. The TER also summarizes the factors that are important in assessing com- pliance with 10 CFR Part 61, Subpart C. It notes on page 90, “some of the assumptions made in the analysis, if incorrect, could lead to noncompliance with the performance objectives.” Current Status Cementitious grouts and related materials are routinely used in the construction industry for a wide variety of applications, some of which closely match EM’s needs. Where the project requirements are the same as or similar to these routine applications, the DOE can simply use exist- ing technology. For example, controlled low-strength materials are used in bulk to fill utility trenches and provide some load-carrying capacity, which would be similar to bulk-filling a large waste tank to provide structural stability. However, there are some requirements that are unique to DOE applications: • Grout mixtures must be suitable for pumping into the tanks, typi- cally through “tremies” (long, movable pipes that allow placement of the grout into specified locations in the tank (Figure 2.8) without the compo- nents of the grout (mainly cement, sand, and water) separating; • They must provide near- and long-term chemical conditions (high pH and low Eh) to maintain the radionuclides and toxic heavy metals in their least mobile forms; and • They must minimize the flow of water through the material (and the consequent leaching of radionuclides and metals from the grout) (NRC 2006b). In recent years the construction industry has begun to build structures with design lives of 75 to 100 years based on a combination of experience with concrete structures over decades and modeling. However, the design lives of grouted DOE wastes are intended to be 10 times longer or more. The short-term behavior of radionuclides and toxic heavy metals in waste- form grouts is reasonably well understood, but the long-term behavior is 

66 ADVICE ON THE DOE’S CLEANUP TECHNOLOGY ROADMAP FIGURE 2.8  Tremie being used to emplace grout in a tank at the INL site. The use of tremies (long flexible pipes) is common in 2-8construction industry. For EM’s ap- Fig the plications, the components of specially formulated grout mixtures must not separate bitmap image before they reach their in-tank destination and solidify. SOURCE: Department of Energy. not. For example, it is well known that the pH of cementitious materials decreases over time due to carbonation (reaction with carbon dioxide from the air), while little is known about how other properties such as Eh change with time. The models currently being used to predict long-term performance of tank grouts and saltstone necessarily extrapolate from very limited and relatively short-term data. Concrete vaults are constructed to contain SRS saltstone. The vault wall, floor, and ceiling are part of the engineered barrier system expected to reduce the ingress and egress of water. Some of the vaults have already cracked and may be transmitting water, but the causes and potential fixes are not yet understood (USNRC 2008). The USNRC outlined several cooperative research efforts focused on the long-term behavior of cementitious systems (Kock 2008). The DOE ce- ment consortium includes both a simulation component and an experimen- tal component. It is led by the DOE and includes the USNRC, Vanderbilt 

PRINCIPAL SCIENCE AND TECHNOLOGY GAPS 67 University, the National Institute of Standards and Technology, Simco, the Netherlands Energy Research Center, and SRNL. It is funded by the USNRC Office of Research. In addition, the USNRC is sponsoring contrac- tor research on cement-based materials, including degradation mechanisms, modeling, fast pathways, hydraulic conductivity studies, and a test bed. The USNRC Office of Research is also examining test methods, designs, addi- tives, and monitoring techniques. Approaches to Bridge the Gap According to this committee’s assessment of information it received, the following approaches have promise to lead to improved understanding of the long-term performance of cementitious materials and, thus, to improve- ments in the materials per se: 1. Improved data to support performance assessment models. The models currently being used to predict long-term performance necessarily extrapolate from very limited short-term data. Current models are believed but not known to be conservative. Monitoring the near- and long-term per- formance in the field would greatly improve the accuracy of the models and allow for adjustments to the grout formulations for future tank closures. The USNRC TER (p. 78) provides an example of how empirical data from the field could be used for this purpose: One of the key elements of DOE’s PA is the [chemical] reduction of Tc- 99 in the wasteform by the addition of slag. As previously discussed, the sensitivity analyses demonstrate quite clearly that the rate and extent of oxidation of the wasteform is a key factor in meeting the protection of the public performance objective. DOE has performed basic research to evalu- ate whether the slag would result in Tc-99 being contained in a reduced form, and installed field-scale saltstone lysimeter tests with and without slag (Cook and Fowler, 1992). . . . Currently, DOE’s estimates for the amount of oxidation of the saltstone over 10,000 years are based primarily on numerical modeling results. It may be possible to exhume and charac- terize a saltstone lysimeter. The depth of the penetration of the oxidation front should be able to be estimated and it would provide excellent model support for a key element of DOE’s PA [italics added]. The USNRC TER also noted inconsistent results for the measured hy- draulic conductivities (permeabilities) of saltstone samples made or tested under different conditions. Very low conductivity must be achieved and maintained for saltstone to meet its performance requirements. Develop- ing improved and quality-assured methods to measure hydraulic conduc- tivities of very low permeability materials can assist the SRS tank closure program. 

68 ADVICE ON THE DOE’S CLEANUP TECHNOLOGY ROADMAP 2. Basic understanding of the chemistry for improving the long-term performance of cementitious waste forms. In addition to limiting access of water to the waste, cement-based materials are expected to maintain the wastes in a chemically reducing environment to reduce the solubility of waste components. Properties such as reducing capability are not relevant in the construction industry; therefore, fundamental research, along with gathering the empirical data described in the USNRC TER above, are spe- cific opportunities for EM to address this knowledge gap. Pore fluids in grouted wastes can concentrate and release contaminants. However, relating the waste inventory in the saltstone to pore fluid concen- tration is complicated by various processes, such as precipitation/dissolu- tion reactions, aqueous complex formation, and sorption. These processes are poorly understood and difficult to quantify, but they must be under- stood in order to ensure that grouted wastes meet long-term performance objectives. 3. Technology development for pipe grouting. The various tank farms have underground piping to carry wastes from one tank to another. Exca- vating these pipes and fittings is not practical, so they will be left in place. They will most likely be grouted with a material similar to that used on the bottoms of the tanks, that is, an engineered grout with reducing properties. This material will need to flow into place under pressure, fill or nearly fill the pipe, and not set until it has done so. Once the material sets, there will be no possibility of pumping additional material into or through the pipe. For this application, the shrinkage of the grout must be controlled. All cementitious materials shrink to some degree on hydration, which would leave the pipe less than completely filled. Materials can be engineered to shrink less, and shrinkage-compensating materials can also be designed. These formulations expand first and then shrink. When the initial ex- pansion is restricted, as it would be inside a pipe, it could induce radial (bursting) stresses in the pipe. Learning how to better control expansion or shrinkage of cementitious materials in confined areas is an opportunity for EM to partner with industry. 4. Improved understanding of crack formation and mitigation in con- crete vaults. Cracking of construction concretes is a well-known phenom- enon. Many of the causes for cracking and the means for mitigation are well understood. However, in construction, cracks are expected and tolerated within certain limits, and cracks exceeding these limits can often be satisfac- torily repaired. New methods of design, detailing, and construction may be required where tighter tolerances or extended durability are necessary. New repair materials and methods suitable for areas with limited accessibility could also be helpful. 

PRINCIPAL SCIENCE AND TECHNOLOGY GAPS 69 Prioritization of the Gap Relative to other science and technology gaps discussed in this section the committee judged the priority of addressing this gap as High. Relative Rating Criteria High Medium Low Volume of waste affected X Potential to reduce technical uncertainty X Potential to affect cleanup schedule X Potential to affect cost X DEACTIVATION AND DECOMMISSIONING The deactivation and decommissioning and facility engineering pro- gram area of the EM roadmap deals primarily with facility characterization; deactivation, decommissioning, and demolition; and closure.24 Principal science and technology gaps the committee identified in this program area are described in this section. Facilities requiring D&D throughout the DOE complex include repro- cessing plants, large production and smaller test reactors, fuel fabrication facilities, gaseous diffusion plants, and laboratories with hot cells—includ- ing all of these facilities’ support structures that typically contain ancillary equipment, piping, and ductwork. In many cases there are complicating factors including poor (and continually degrading) condition of structures, associated chemical hazards, and nearby active facilities with ongoing operations. While hundreds of DOE facilities have undergone D&D, some 3,000 remain to be decontaminated and removed or closed, including many of the most challenging ones.25 A previous report (NRC 2001c) noted that cleanup of facilities will be technically challenging due to (i) personnel hazards; (ii) large size of facilities, including those with massive shielding structures; (iii) complex, crowded and often retrofitted arrangement of equipment and support structures; (iv) poorly understood and difficult-to- characterize contaminants; and (v) lack of decisions on end states. Many buildings and facilities are to be partially or completely demolished while some massive structures will be decontaminated and left in place. The con- taminants to be removed include solids and liquids that can be radioactive, 24 For convenience this report will use the abbreviation D&D in referring to EM work in this program area. 25 See http://www.em.doe.gov/Pages/BudgetPerformance.aspx. 

70 ADVICE ON THE DOE’S CLEANUP TECHNOLOGY ROADMAP chemically hazardous, or both. Some contaminants may be easy to remove and others strongly bound to a substrate (e.g., concrete, steel). According to information gathered by the committee, the most difficult D&D challenges include radiochemical separation facilities at Hanford, Idaho, and SRS; production reactors at SRS; gaseous diffusion plants at Oak Ridge, Paducah, and Portsmouth, plutonium processing plants at Hanford, Los Alamos, and SRS; tritium processing facilities at SRS (NRC 2001c), and support facilities (including sewage lines) at SRS (Whitaker 2008). DD-1: D&D work relies on manual labor for facility characterization, equipment removal, and dismantlement. Currently D&D projects require extensive hands-on, manual labor that unavoidably exposes workers to hazardous conditions (Figure 2.9). Besides the rather obvious hazards to workers who manually dismantle, size reduce (cut up), and remove contaminated structures and equipment, each facility requires extensive characterizations to determine the nature of contaminants before, during, and after D&D. Characterization exposes workers to radiation and other hazards and is costly, amounting to some 15 to 25 percent of overall D&D budgets (NRC 2001c). Work must sometimes be done in high-radiation environments. For example, at Idaho a techni- cal challenge is to characterize and remove contamination in pipelines and other structures that have high-radiation fields (up to 1,600 rads/hour) and are located under a building at the site (NRC 2007c, p. 28). Workshop panelists representing Oak Ridge agreed that D&D is a top priority for the site, mainly due to challenges presented by the gaseous diffusion plants (manual removal of transite siding from these very large buildings was cited (Figure 2.10) and other deteriorating structures (NRC 2007c; McCracken 2007). SRS D&D priorities are worker protection and characterization of facility “hot spots” (NRC 2007c, p. 27). Impact of the Gap Safety of workers and of the public is the primary consideration in the EM cleanup. Worker safety is a criterion for contractors. Should an incident occur that harms a worker or could have caused harm, operations are halted until the incident has been thoroughly investigated, the cause is determined, and measures to prevent such future incidents are implemented. No matter how carefully planned and carried out, hands-on D&D work carries a high risk for radiation exposure, bodily uptake of radioactive or hazards materials, and injury. 

PRINCIPAL SCIENCE AND TECHNOLOGY GAPS 71 FIGURE 2.9  Hands-on D&D work. Facility D&D often requires hands-on work with large, contaminated equipment in hot, confined spaces. Although uncomfort- Fig 2-9 able, personal protective equipment like that worn by the worker in this photograph bitmap image is necessary to protect workers from the uptake (skin, mouth, nose) of radioactive or other hazardous substances and from physical hazards. SOURCE: Department of Energy. FIGURE 2.10  Transite removal at Oak Ridge. Transite was a commonly used siding material throughout the DOE complex. Today’s workers must wear personal pro- tective equipment and follow special procedures to remove this asbestos-containing siding. Transite is heavy and often has to be handled in confined, elevated work spaces as shown here. Fig 2-10 SOURCE: Department of Energy. bitmap image 

72 ADVICE ON THE DOE’S CLEANUP TECHNOLOGY ROADMAP Current Status Manual labor has been key to EM’s D&D work, including the suc- cessful closure of the Rocky Flats site under budget and ahead of schedule. Rocky Flats was formerly a major plutonium-handling site, which has now been converted into a wildlife refuge—a major accomplishment. Hands-on labor for D&D is a good example of the committee’s considering technol- ogy gaps as potholes in a road that EM can work around. EM and its contractors can and have managed worker safety for hands-on D&D. Nonetheless, R&D toward removing workers from a hazardous environ- ment could provide a better solution. Robotics and remote manipulation for sensing, inspection, measure- ment, and tank waste remediation have been developed and deployed to some extent at both the Savannah River and Hanford sites. DOE has made limited use of some robotic technology as part of the Glovebox Excavator Method used to demonstrate retrieval of buried TRU waste at Idaho (NRC 2005, p. 43). Researchers at INL have been exploring the possibility of using semi- autonomous robotic systems for detection and characterization in radiolog- ical environments. These systems may reduce some uncertainties inherent in different training and skill levels among operators while allowing tasks to be completed more quickly than in the case of purely teleoperated systems (Nielsen et al. 2008). In all cases, the purpose of employing robotic and remote systems is to reduce D&D worker risks while accelerating the pace and accuracy of the remediation operation. SRNL is extending its previous experience with remote devices for use in radiation areas to develop robotic and teleoperated systems for home- land security and defense applications. Non-DOE agencies and universities, including the National Aeronautics and Space Administration (NASA), National Oceanic and Atmospheric Administration, and Carnegie Mellon University, conduct research on robotics and remote-operator systems for the Department of Defense and for ocean and extraterrestrial exploration. There are also recent efforts outside the United States to develop robot- ics and remote systems for decommissioning of former nuclear power facili- ties. For example, a group at Lancaster University in the United Kingdom has been funded by the Nuclear Decommissioning Authority to develop a multiarmed robotics system that would allow D&D operations in the United Kingdom to be faster, safer, and more cost-effective, and reduce the radioactivity dose levels to which workers are exposed (Bakari et al. 2007). Work at the French Atomic Energy Agency and COGEMA has focused on radiation-hardened electronics and force feedback mechanisms used in telerobotics operations involving spent fuel (Desbats et al. 2004). 

PRINCIPAL SCIENCE AND TECHNOLOGY GAPS 73 Approaches to Bridge the Gap According to this committee’s assessment of information it received, the following have promise for future EM R&D: 1. Improved technologies that could reduce worker exposure by re- ducing the need for manual sample collection. These include: • Devices for rapid characterization of low levels of contami- nation (radionuclides and EPA-listed substances) on surfaces of con- struction materials and equipment, including devices that can detect very-low-energy beta emitters (e.g., tritium), low-energy photon emit- ters (iodine-129), and beryllium; • Minimally invasive methods to characterize contaminant con- centrations as a function of depth in construction materials, especially concrete; and • Instruments for remote mapping of radionuclide contamina- tion at low levels that can differentiate specific radionuclides, including beta and alpha emitters. 2. Greater use of robotics to reduce manual labor and worker risks. NRC (2002) recommended that DOE develop robotic technologies for retrieval and repackaging of buried waste. NRC (2001c) recommended research to develop intelligent and adaptable robotic systems that can be used for facility decommissioning. Next-generation robotic systems will need to be: • Adaptable to a variety of environments and topographies; • Semi-autonomous to provide a more intuitive human-robot interface, prevent accidents, and optimize execution of tasks; and • Highly reliable. Such needs were recognized in EM’s former D&D Focus Area 10 years ago (Staubly and Kothari 1998) and remain at the forefront of R&D in robotics. Prioritization of the Gap Relative to other science and technology gaps discussed in this section the committee judged the priority of addressing this gap as High. 

74 ADVICE ON THE DOE’S CLEANUP TECHNOLOGY ROADMAP Relative Rating Criteria High Medium Low Volume of waste affected X Potential to reduce technical uncertaintya X Potential to affect cleanup schedule X Potential to affect cost X aIncluding risks to workers. DD 2: Personal protective equipment tends to be heavy and hot and limits movement of workers. As described in DD-1, manual D&D work at all sites requires workers to perform safely and efficiently in hazardous environments. Broadly speak- ing, personal protective equipment (PPE) can range from standard items such as coveralls, safety glasses, and gloves, to face masks with capability to filter or detoxify airborne contamination (“assault masks”), to full-body anti-contamination suits for work in heavily contaminated areas (Figure 2.9). Anticontamination suits encapsulate the entire body in an impervious suit, and provide safe breathing air by means such as filtration of ambi- ent air, use of self-contained breathing apparatus, or an external supply of uncontaminated air delivered through a flexible hose. PPE for less-contaminated workspaces consists of some type of protec- tive clothing, often in multiple layers, which encloses most or all of the body. PPE is often heavy and bulky, resulting in limitation of motion, extra exertion, and overheating with the consequent risk of heat stress (Bernard 1999). Protective clothing that does not allow perspiration to escape in- creases body temperature, which reduces worker comfort and productivity (DOE 1998b). Impact of the Gap The limitation of motion and extra exertion imposed by PPE required in high-contamination zones can cause worker stress and reduce the effi- ciency of D&D work. PPE with externally supplied cool air can reduce heat stress but can have various limitations and problems related to the supply hose. During its Idaho visit, the committee was shown a waste retrieval operation in the Radioactive Waste Management Complex in which work- ers can operate excavation equipment for only short periods of time due to the risk of heat stress. This, coupled with the time to don and doff PPE, increases the duration and cost of D&D activities. 

PRINCIPAL SCIENCE AND TECHNOLOGY GAPS 75 Current Status PPE is used throughout the nuclear and hazardous materials industries. There are companies that develop and manufacture PPE (see, e.g., Frham Safety26 and G/O Corporation27). EPRI and USNRC are supporting tech- nology for improving PPE. Shedrow (2008) noted that SRNL has developed a variety of PPE technologies that have been used in environmental reme- diation work at SRS and other locations. Approaches to Bridge the Gap According to this committee’s assessment of information it received, there is a need for PPE designed for elevated temperatures and longer expo- sures in contaminated environments. Lighter and cooler PPE would allow workers to safely remain longer in the presence of hazardous materials. There are opportunities to adapt available technologies (e.g., from NASA, Department of Defense). For example, adaptations of NASA protective clothing technology have been examined for use in development of protec- tive clothing for firefighters (Foley et al. 1999). The Department of Defense has supported a number of programs for development of advanced imper- meable “NBC” (nuclear/biological/chemical) anticontamination clothing for a number of years, citing this area of need in the Defense Technology Area Plan (DOD 1999). This technology has not been adapted and adopted in D&D applications. Further evaluation would seem appropriate. Robotics and remote or teleoperated techniques will also limit worker exposure, although there are circumstances (i.e., inspection, removal in very complex areas, sensitive structures) where manual labor is essential. Prioritization of the Gap Relative to other science and technology gaps discussed in this section the committee judged the priority of addressing this gap as Low. Relative Rating Criteria High Medium Low Volume of waste affected X Potential to reduce technical uncertaintya X Potential to affect cleanup schedule X Potential to affect cost X aIncluding risks to workers in this instance. 26  ttp://frhamsafety.com/anti-c/encapsulating_suit.htm. h 27  ttp://www.gocorp.com/. h 

76 ADVICE ON THE DOE’S CLEANUP TECHNOLOGY ROADMAP DD-3: Removing contamination from building walls, other surfaces, and equipment can be slow and ineffective. Decontamination of facilities and equipment is carried out at multiple stages of the decommissioning process in order to lower worker exposure, prepare equipment for disassembly and removal, and prepare a facility for tear-down and removal (to limit release of contaminants prior to further treatment and disposition of the debris). A primary objective of decon- tamination procedures is to generate a small volume of the most hazardous waste, while the larger volumes of waste have low or no hazard, thus re- ducing the cost and long-term risk of their disposal. Some decontaminated equipment or facilities might be recycled or reused. The end state of any decontamination activity must be consistent with both site-specific and overall DOE cleanup objectives. Concrete, such as that in the large canyon buildings on the SRS and Hanford sites and reactor shielding structures at multiple DOE sites, consti- tutes most of the volume and weight (estimated at over 27 million tons) of DOE’s surplus facilities. Because of its inherent porosity, its heterogeneous surface structure (pits, cracks, and smooth and rough areas on both the macro- and microscopic scales), and its chemistry, concrete poses special challenges for decontamination. At present, the usual method for removing surface contamination is called “scabbling”—the physical removal of the surface by workers in pro- tective clothing using power tools. This procedure generates a great deal of dust and is hazardous to workers. Because of long-term exposure, the concrete is often contaminated to a depth of several millimeters beneath its surface (DOE 2000), and in some cases, such as for tritium, consider- ably deeper. In many instances, paints, sealers, and varnishes on concrete surfaces create a laminate problem, with aged materials being harder to decontaminate than more recent deposition (NRC 2001c). Contaminated equipment including glove boxes, shielded cell liners, lead shielding, and plastic parts, along with heavily corroded surfaces, pose particular problems due to geometries and occluded structures that trap contaminants. In addition, effectiveness of D&D methodologies can be severely compromised due to the inherent difficulty of characterizing both the chemical nature of contaminants and the degree of their removal follow- ing decontamination resulting from occluded, porous, and heterogeneous surfaces of degraded structural building materials (Halada 2006). Before, during, and after the process of decontamination, it is necessary to identify contaminants on concrete and other structural surfaces. Nondestructive methods would be far preferable to the physical removal of samples (e.g., cement cores, metal coupons) for analysis. 

PRINCIPAL SCIENCE AND TECHNOLOGY GAPS 77 Impact of the Gap Current decontamination processes used by D&D contractors are labor-intensive and costly, and there is the ever-present risk of exposure to toxic and radioactive materials; see DD-1. These processes also generate large volumes of contaminated secondary wastes and often leave behind unwanted residual contamination. The risk of accidents is increased by the bulky protective clothing; see DD-2. Because of cost and hazards, cleanup contractors often choose to dispose of contaminated equipment and con- struction materials as wastes rather than to decontaminate and recycle them. While current baseline decontamination technologies probably can be made to work for future D&D work, there are opportunities to do the job more safely and cheaply and achieve higher degrees of decontamination by developing and using new technologies. Current Status The EPA has recently conducted two workshops on decontamination methods for chemical, radiological, and biological contaminants through its Office of Research and Development’s National Homeland Security Center (EPA 2005, 2006a). In addition, the EPA has developed a reference guide, “The Technology Reference Guide for Radiologically Contaminated Sur- faces,” which provides a broad overview of chemical and physical methods for removing contamination from surfaces (EPA 2006b). These surveys and associated reports consider a broad range of options for decontamination. The technological challenges considered in the EPA report have much in common with DOE site needs, including a need for faster and more effec- tive decontamination methods, determining surface chemistry interactions, difficulties with vertical surfaces and reaching high work areas with de- contamination equipment, decontamination of tiny cracks and seemingly inaccessible areas, subsurface effects, and waste generation. Investigators at INL completed a comprehensive study of removal and collection of radioactive contamination from building exteriors, which was supported by the Defense Advanced Research Projects Agency (Demmer at al. 2007). Activities in the United Kingdom and Canada are also of interest. For example, the effect of weathering and other environmental conditions on the association of radiological contamination with porous surfaces and resulting implications for decontamination have been considered in research by the Chemical, Biological, Radiological-Nuclear and Explosives Research and Technology Initiative Secretariat of the Defence Research and Develop- ment Canada, Centre for Security Science.28 28 See http://www.css.drdc-rddc.gc.ca/crti/invest/rd-drt/02_0067rd-eng.asp. 

78 ADVICE ON THE DOE’S CLEANUP TECHNOLOGY ROADMAP Approaches to Bridge the Gap Scientific understanding of the interactions among contaminants and construction materials is fundamental to developing more effective D&D technologies. Such information includes how contaminants bind to steel and concrete surfaces; how they penetrate into these materials; their migra- tion into pores, fissures, and welds; and time-dependent aging effects. NRC (2001c) identified decontamination as an important science and technology gap and recommended specific areas of research needed to improve decon- tamination technologies, including: • Development of a fundamental understanding of the chemical and physical interactions of important contaminants with the primary materials of interest in D&D projects, including concrete, stainless steel, paints, and strippable coatings to gain a better understanding of how contaminants bind to and penetrate these materials. This would involve understanding the interactions both kinetically and thermodynamically under a variety of conditions (pH, temperature, ionic strength); • Development of dry decontamination technologies, including use of supercritical fluids such as carbon dioxide, that can be used to remove high levels of contamination with minimal secondary wastes (Appendix D); • Exploration of the role of nanotechnology (for more efficient che- lating) and biological mechanisms (including bioleaching, biosurfactants, biocatalysis, and cell-less enzymatic processes) for more efficient and rapid decontamination methods; • Advanced methods to leach/migrate contaminants from cementi- tious matrices (Appendix D); and • Development of decision tools for determining optimal decontami- nation approaches. Prioritization of the Gap Relative to other science and technology gaps discussed in this section the committee judged the priority of addressing this gap as Medium. Relative Rating Criteria High Medium Low Volume of waste affected X Potential to reduce technical uncertainty X Potential to affect cleanup schedule X Potential to affect cost X 

PRINCIPAL SCIENCE AND TECHNOLOGY GAPS 79 CONCLUSIONS This chapter has presented 13 gaps that the committee views as the principal impediments to the EM site cleanup program. They are obstacles or impediments in the sense that they can represent likely causes for sched- ule delays, cost increases, and potential failures to meet currently envisaged cleanup objectives. Developed through the committee’s site visits and other information gathering, all of these gaps are worthy of EM’s consideration in developing future science and technology roadmaps. The committee was mindful of the research initiatives set forth in the EM roadmap but has provided its own independent assessments in this chapter. The committee’s prioritization of these gaps, given in Table 2.1, reflects a variety of technical judgments, including schedule and budget impacts, risk reduction, and likelihood of new technology developments that can bridge the gap. The committee has not attempted to be prescriptive by rec- ommending specific research to address each gap, but rather it has indicated R&D approaches that it judges are most likely to bear fruit. The committee used this chapter as a basis for developing the remain- der of this report. Chapter 3 describes the personnel expertise and physical infrastructure that EM will need to carry out this R&D. Chapter 4 de- scribes approaches and opportunities for EM to leverage its R&D work with other organizations.