Tank Wastes Planned for On-Site Disposal at Three Department of Energy Sites: The Savannah River Site: Interim Report (2005)

In the Ronald Reagan National Defense Authorization Act of 2005 (Section 3146 of Public Law 108-375), Congress directed the Department of Energy (DOE) to request a study from the National Academies that evaluates DOE’s plans for managing certain radioactive wastes stored in tanks at its sites in Idaho, South Carolina, and Washington.¹ The wastes addressed in this study are from reprocessing of spent nuclear fuel, exceed certain concentration limits,² and are planned for disposal at the sites mentioned above.

Congress asked the National Academies³ for an interim and a final report addressing this task. According to the Defense Authorization Act, the interim report “shall address any additional actions the Department should consider to ensure that the Department’s plans for the Savannah River Site, including plans for grouting the tanks, will comply with the performance objectives [of 10 CFR 61⁴] in a more effective manner” (Section 3146 (e)(A)). This document fulfills the interim report request.

Congress requested this study at the same time another provision of the same law (Section 3116) provided the basis for DOE, in consultation with the U.S. Nuclear Regulatory Commission (USNRC), to determine that tank wastes at the South Carolina and Idaho sites meeting certain listed criteria are not high-level waste (HLW).⁵ Such wastes may then be disposed of on-site.

The Savannah River Site has 51 underground tanks that are used for storing 138,000 cubic meters (36.4 million gallons) of hazardous and radioactive waste from chemical processing of spent nuclear fuel and related operations.⁶ Tank construction and characteristics vary, but the typical tank is a large cylindrical carbon steel and reinforced concrete structure buried at a shallow depth (1 to 3 meters below the surface). The tanks’ sizes range from about 2,800 cubic meters (m³) to 4,900 m³ (750,000 to 1.3 million gallons). The largest tanks are approximately 26 meters (85 feet) in diameter and 11 meters (35 feet) from the inner tank floor to the center of a domed ceiling. Most of the tanks are equipped with dense networks of vertical and horizontal cooling pipes, referred to as cooling coils (see Figure S-1). These cooling coils are used to remove heat produced by radioactive decay in the waste.

Twenty-seven of the tanks have a full secondary containment (i.e., a tank inside another tank) and are considered “compliant tanks” under the site’s Federal Facility Agreement,⁷ which regulates storage and disposal of hazardous waste at the site. The remaining tanks do not have complete secondary containment and are considered noncompliant. Visual inspections and conductivity probes in the tanks and in the annuli of the tanks have shown that about half of the noncompliant tanks have leaked in the past (although the leaks were confined to the tank’s annulus in all but one case).

FIGURE S-1 Photograph of the interior of a Type I tank (Tank 4) prior to receipt of wastes. SOURCE: Caldwell (2005a).

Although the composition of waste in each tank varies, the tanks generally contain a bottom layer of a peanut-butter-like deposit of insoluble solids (referred to as sludge), a layer of crystalline solids (the saltcake), and a salt solution (the supernate). The term “salt waste” is sometimes used to refer to saltcake and supernate. Although the sludge represents less than 10 percent of the volume, it contains about half of the radioactivity in the waste tanks,⁸ mainly from insoluble actinides and strontium salts. The other half of the radioactivity is mostly in the supernate, where the soluble radionuclides, mainly cesium-137, are in solution. A fraction of the soluble radionuclides is also trapped as liquid in the interstices of the saltcake.

DOE has argued that it is impractical to dismantle and remove the tanks after the waste has been retrieved because of the exposures incurred by workers from radioactive residues and because of the overall prohibitive costs of exhuming such large structures. The committee has not seen analyses to support this claim. For each tank, the general plan is to retrieve the bulk of the waste, clean up the tank to the “maximum extent practical,”⁹ and close the tank in place, according to milestones agreed to in the site’s Federal Facility Agreement. Because of practical limitations on waste retrieval, “emptied” tanks will still contain variable amounts of the radioactive waste (the “heel”), depending on the success of the retrieval and cleanup process.

DOE plans to close emptied tanks by placing layers of engineered grout to encapsulate and stabilize the tank heel and a controlled low-strength material to provide structural support against tank collapse and act as a physical barrier that inhibits the flow of water through the residual waste. Tanks that do not have a concrete roof would have a high-strength layer of grout that would serve as an intruder barrier. An engineered cover to retard infiltration to the tanks after closure is also under consideration.

DOE’s plan to manage the bulk of the waste retrieved from the tanks is to separate the radioactive from the nonradioactive components, the latter of which make up most of the waste volume. This processing generates two waste streams: (1) a high-activity waste stream, which will be immobilized and disposed off-site in a high-level waste repository,¹⁰ and (2) a low-activity waste stream, which is to be disposed on-site.

At the Savannah River Site, DOE already retrieves sludge and then processes and immobilizes it in glass at its Defense Waste Processing Facility (DWPF). These operations generate as a secondary product a relatively low-activity liquid waste, referred to as the DWPF recycle stream, which is returned to the HLW tanks. To separate highly radioactive constituents of the salt waste, DOE proposes to utilize three different processes¹¹ that will be available at different times and have different capabilities. Two low-capacity processes are expected to be available sooner and are referred to as “interim” processing by DOE. These are the deliquification, dissolution, and adjustment (DDA) process, which could begin

immediately upon approval of the waste determination by the Secretary of Energy in accordance with Section 3116 of the 2005 National Defense Authorization Act, consultation with the USNRC, and permitting by the state of South Carolina; and the actinide removal, modular caustic-side solvent extraction process (ARP/MCU), which is expected to begin operations in 2007. A high-capacity chemical processing facility, called the Salt Waste Processing Facility, is scheduled to be available in 2009 and could be supplemented by the ARP, if needed.

DOE indicated to the committee that the Savannah River Site is facing a “tank space crisis” because of net waste inputs from current waste processing and waste removal operations. To alleviate the tank space crisis, DOE is proposing to begin processing salt waste using DDA as soon as possible (see Figure S-2). The low-activity waste streams from these three processes will have varying concentrations of radioactivity and will be mixed with cementitious material to form “saltstone” and disposed on-site as a monolith in near-surface concrete vaults.

Although DOE, its regulators, and others worked with the committee to provide the information needed for this study, some data were not available (not yet collected, not yet generated, or not yet made public), and some plans had not yet been formulated or finalized

FIGURE S-2 Waste flows in the Savannah River Site waste management plans. Note that the sizes do not necessarily scale with the sizes of the waste flows.

when this report was written.¹²Appendix B describes the main documents to which the committee had access and the missing pieces of information to assess DOE plans for compliance with the performance objectives set forth in 10 CFR 61.

Therefore, the committee was unable to evaluate fully what, if any, actions are needed for DOE to comply with these performance objectives. However, the committee was able to evaluate factors that reduce risk and recommends actions to (1) reduce the waste left on-site and (2) increase DOE’s understanding of the long-term performance of waste forms and other barriers to the release of radionuclides. These actions will increase confidence in DOE’s ability to comply with the performance objectives in general and conform with the requirement to take actions to make releases of radioactivity to the environment as low as reasonably achievable (ALARA), with economic and social considerations taken into account. Findings and recommendations address four major issues: (1) near-term and long-term risks; (2) the tank space crisis; (3) Class C limits and performance objectives; and (4) research and development needs. The following findings and recommendations are based on the information available to the committee at the time of writing this interim report and may be extended in the committee’s the final report.

Finding 1a: By far the greatest reductions in near-term probability and quantity of radionuclide and hazardous chemical releases to the environment are achieved by bulk removal and immobilization of liquid, salt, and sludge from the noncompliant high-level waste tanks. The tank heels that remain after bulk removal contain a smaller quantity of waste that is less mobile and constitutes a much lower near-term probability of release.

Finding 1b: The Savannah River Site Federal Facility Agreement has schedules for waste removal from and closure of the noncompliant tanks. For some tanks, the tank-closure step immediately follows the waste-removal step, making them appear to be coupled. This coupling could limit the time available for tank-waste removal and consequently could determine how much waste can be removed to “the maximum extent practical.” A decoupled schedule is already planned for a limited number of tanks, as shown in Appendix F. Decoupling allows the consideration of a wider set of options for removing and/or immobilizing residual waste (especially for tanks that have significant obstructions that complicate waste removal), which could reduce long-term risks.

Recommendation 1: DOE should decouple tank waste removal and tank closure actions on a case-by-case basis where there are indications that near-term (5-10 year) techniques could become available to remove tank heels more effectively, safely, or at a lower cost. In evaluating schedules for each tank, DOE should consider the risks from postponing tank closure compared with the risk reductions that could be

achieved if the postponement improves heel removal. Although the committee believes that postponing tank closure need not extend the closure dates of the tank farms, DOE should work with the State of South Carolina to revise the schedule for closure of a limited number of the tanks that contain significant heels, if necessary.

The committee agrees with DOE’s and South Carolina’s overall approach to cleanup at the Savannah River Site: bulk removal of the waste containing the majority of the mobile radionuclides is the highest priority to reduce release of radioactive materials to the environment in the near term. The noncompliant tanks, about half of which have a history of leakage, demand attention first, but nearly all of the tanks are beyond their design lifetimes.

Filling a tank with grout is, from a practical point of view, an irreversible action, although it is conceivable to open a tank and excavate the grout if absolutely necessary. Moreover, postponing closure of some tanks for several years would appear to have essentially no effect on near- or long-term risk. The current approach of coupling cleanup and closure schedules forecloses options that may become available in the near future (e.g., using alternative technologies to reduce the radioactive heel [source] and/or using other types of immobilizing material to fill the tank).

DOE should decouple cleanup and closure schedules, keep as many options open as practical, and regularly assess technology developments and alternatives to reduce long-term risks presented by the tank heels. DOE should make additional investments in research and development to enhance tank waste retrieval (reducing the source term), improve residual waste immobilization (stabilizing the source term), or reduce the ingress of water once the tanks are closed (protect the source term), as stated in Recommendation 4. In some cases, tank closure need not be delayed, such as in tanks that have small heels (i.e., as small as the heels in Tanks 16, 17, and 20) and/or low concentrations of radionuclides, or if risks specific to the tank require early closure (i.e., as soon as waste removal is completed). Conversely, delaying closure may be warranted for tanks with large heels or high concentrations of radionuclides. This approach need not necessarily affect the final closure date of the tank farm, which will occur later than 2022, the milestone for closure of the noncompliant tanks. If new technologies become available in the near future (i.e., 5-10 years), it may be possible to clean up and/or close tanks faster (possibly leaving less waste behind), thus meeting the final milestone for the tank farms.

As DOE considers delaying closure for some tanks, it has to evaluate the advantages and disadvantages from both a risk and a cost perspective. If DOE can relax other constraints on tank waste removal, such as the tank space problem, delaying tank closure could free up funds planned for closure activities, and those funds could be devoted to enhancing waste removal, waste processing, and confidence in the near- and long-term performance of the waste immobilization and tank fill materials. Similarly, research and development require funds, but if they are successful they could result in lower costs and increased safety overall (see Finding and Recommendation 4).

Finding 2a: The lack of compliant tank space does appear to be a major problem because of continuing waste inputs and the anticipated future needs for space to support site operations and tank cleanup. As presently operated, sludge waste processing results in a net addition of waste to the compliant tanks. Salt waste processing will also require storage volume in compliant tanks for batch preparation and other operations.

Finding 2b: DOE plans to use the deliquification, dissolution, and adjustment process to free up space in compliant tanks. While DOE analyses so far suggest that the wastes from this process would meet the performance objectives in 10 CFR 61, it achieves less radionuclide separation than other planned processes. While waste from the DDA process represents only 8 percent of the volume of low-activity waste to be generated during salt waste processing, it contains 80-90 percent of the radioactivity that is projected to be sent to the Saltstone Disposal Vaults.

Recommendation 2: DOE and other involved parties should consider options other than DDA to alleviate the impending crisis in usable storage in compliant tanks. Options include actions that (1) reduce waste inputs to the tanks, such as redirecting the DWPF recycle stream for disposition in the Saltstone Facility; and (2) actions that free up usable volume in compliant tanks, such as using noncompliant tanks not known to have leaked for emergency storage volume.

Waste retrieval, processing, and tank cleaning operations continuously add secondary wastes to the tanks; in addition, space in compliant tanks is needed to prepare feeds for the high-level and salt waste processing facilities. Moreover, DOE is maintaining the equivalent of a full tank capacity—4,900 m³ (1.3 million gallons)—in empty compliant space for emergency purposes at all times. Hence the “tank space crisis.”

DOE plans to address the tank space problem in the short term by implementing the DDA process. This process uses physical rather than chemical means to accomplish cesium separation (i.e., draining interstitial liquid present in the saltcake and then dissolving the saltcake and grouting it into saltstone (see Figure S-2).¹³ The saltstone from this process is expected to contain cesium concentrations that are two orders of magnitude higher than the waste from the chemical processes that eventually will be used in the Salt Waste Processing Facility (albeit still considerably lower than Class C limits). Even these higher levels of cesium may not cause projected doses from the Saltstone Vaults to exceed dose limits, although as noted earlier, details underlying a performance assessment for DDA saltstone were not available for committee examination. However, this raises the following question: Does this process remove radionuclides to the maximum extent practical?

The tank space crisis forces DOE to engage in increasingly complex operations to ensure that there is sufficient space to continue waste processing. Hence, the tank space crisis may increase the possibility of accidental worker exposure to radiation, the chance of operational accidents, and the chance of waste leakage during transfers. In its recommendation, the committee suggests alternative options to DDA to mitigate the tank space crisis.

Finding 3: The future site-specific risks posed by wastes disposed of on-site is the primary issue of concern in this study. Such risks are determined by the radionuclide and chemical quantities and concentrations, their conditioning, their interactions with the environment, and their bioavailability, not by the relationship of radionuclide concentrations to generic limits such as those for Class C low-level waste. The National Defense Authorization Act Section 3116 requires the use of the performance objectives in 10 CFR 61 to limit and minimize these risks.

Recommendation 3: When deciding what wastes may be disposed of on-site, DOE and other involved parties should ensure that discussions focus on how radionuclide and chemical quantities and concentrations, their conditioning, their interactions with the environment, and their bioavailability affect site-specific risk.

The Class C limits are not a criterion for acceptability of on-site disposal of tank wastes from reprocessing of spent nuclear fuel under the present law but are sometimes discussed as if they were. The Class C limits were developed for a diverse commercial sector to establish limits on what is generally acceptable for near-surface disposal, based in part on assumptions about the overall set of wastes destined for disposal. According to Section 3116, comparison of radionuclide concentrations in waste to Class C limits is relevant to waste disposition decisions only procedurally, in that DOE must develop its disposal plans in consultation with USNRC.

Rather than Class C limits, site-specific risk assessments are the bases for determining whether the facility meets the performance objectives in the regulations. These risks depend on radionuclide quantities and concentrations, their conditioning, and their interactions with the environment.¹⁴ The performance objectives and waste acceptance criteria constrain the overall quantity of radioactive material that can be disposed in a facility.¹⁵

Acceptable radionuclide concentrations (and/or inventories) and distributions should be determined as a result of a properly constituted and implemented risk assessment¹⁶ that takes into account measured and/or projected radionuclide concentrations, spatial variability of the concentrations, and attendant uncertainties. Such a risk assessment was not available at the time of report writing (see Appendix B).

Congress recognized the importance of the performance objectives for evaluating site-specific near-surface disposal of waste in Section 3116 of the 2005 National Defense Authorization Act by explicitly including these objectives as the basis for determining whether waste is HLW instead of relying on the radionuclide concentrations that define the upper boundary of Class C waste. All substantive technical criteria that DOE’s determination must meet (e.g., performance objectives, remove highly radioactive radionuclides to the maximum extent practicable) apply irrespective of whether a waste is less than or greater than Class C.

Finding 4: Focused research and development could help DOE reduce the amount, improve the immobilization, and test some of the assumptions used in performance assessment of tank waste to be disposed of at the Savannah River Site. These actions could reduce the risks to humans and the environment and improve confidence in DOE’s risk estimates. These research and development activities could

also increase DOE’s ability to demonstrate compliance with the performance objectives in 10 CFR 61.

Recommendation 4: DOE should fund research and development efforts focused on providing deployable results within 5-10 years on the following topics: (1) in-tank and downstream processing consequences of chemical tank-cleaning options, (2) technologies to assist in tank-waste removal, including robotic devices, and (3) studies of the projected near- and long-term performance of tank-fill materials such as grout.

To reduce long-term risks to the site and test the assumptions in the performance assessment, the committee recommends that DOE perform focused research and development to enhance tank waste retrieval and residual waste immobilization. Tank waste retrieval could be enhanced using better mechanical or chemical tools. Tank waste retrieval is currently performed using hydraulic technologies (i.e., water jets) and, to a certain extent, robotic devices and chemical cleaning agents (i.e., oxalic acid). The committee believes that additional research and development on mechanical tools, including but not limited to robotic devices and chemical cleaning could reduce the tank heels, especially in tanks with cooling coils. DOE should further evaluate the effectiveness of residual waste immobilization by conducting durability studies of grout (and alternative fill materials).

These activities may increase confidence in DOE’s management plans or may cause DOE to revise some of the assumptions used in the performance assessment. Testing assumptions and improving DOE’s knowledge base might increase its ability to comply with the performance objectives specified in the law. Research and development activities should be limited to those technologies that are promising and at a near-deployment stage (i.e., they could provide results within 5 to 10 years, in time to be implemented during the tank closure process). All noncompliant tanks are scheduled to be closed by 2022. A technology developed in the next 5-10 years could be deployed in time to address the most challenging tanks (i.e., those with cooling coils).

The committee believes that a nonradioactive test bed for retrieval technologies that can be adapted to simulate a variety of tank situations (i.e., recalcitrant heels, cooling coils, debris) should be maintained. The Pump Test Tank, a partial Type IV tank mockup at the mostly decommissioned TNX facility used for testing and equipment before deployment, and similar test beds at other sites, are candidates for this role. The Hanford Site also has a mockup of a single-shell tank used for similar purposes. The committee will further address the need for experimental retrieval facilities in its final report.

The committee’s full task is to review and evaluate DOE’s plans to manage radioactive waste streams from reprocessed spent fuel that exceed the Class C concentration limits and are planned for on-site disposal at the Savannah River Site, the Idaho National Engineering and Environmental Laboratory, and the Hanford Reservation. Congress requested assessments of the following: DOE’s knowledge of the characteristics of the wastes; additional actions DOE should take in managing these wastes to comply with the performance objectives; monitoring plans; existing technologies and technology gaps for waste management; and any other matters that the committee considers appropriate and directly relevant. For its interim report, the committee was charged to examine whether DOE’s plans to manage its radioactive waste streams at the Savannah River Site will comply with the performance objectives of 10 CFR 61.

Compliance with the performance objectives depends upon the amount of radioactive material left onsite, the manner in which it is immobilized, its interaction with the environment and its interaction with ecological and human receptors. As noted above, some critical data, analyses and plans were not available when this report was written: the performance assessment for closed tanks; plans for residual waste characterization; plans for tank annuli and tank-system piping; support for assumptions, estimated levels of conservatisms, and sensitivity analyses for performance assessment calculations; and long-term monitoring plans are examples of the missing information. In this interim report, the committee has fulfilled the charge to the extent possible by focusing mainly on the amount of waste left in the tanks and in the Saltstone Vaults at the Savannah River Site. The committee has made findings and recommendations on four major issues:

1. near-term and long-term risks in the context of tank waste removal and the schedule for tank closure;

2. the tank space crisis and options to alleviate the crisis;

3. the roles of the Class C limits and the performance objectives in determining whether on-site disposal is acceptable; and

4. research and development needs, particularly in-tank and downstream consequences of chemical cleaning options, technologies to assist in tank waste removal, including robotic devices, and studies of the projected near- and long-term performance of tank fill materials, such as grout.

The committee is still examining the interactions of the tanks and the saltstone with the surrounding environment; the role of environmental monitoring; the role of the point of compliance in meeting the performance objectives; and the role of modeling in the performance assessment. These topics are relevant to all three sites and will be addressed in the final report, along with the rest of the statement of task. For a substantive analysis, the information described above will be needed at all sites. In addition, because the wastes and the site conditions differ, the topics investigated in this report will also be examined at the Hanford and Idaho sites. These investigations at other sites will have an impact on the committee’s views on the Savannah River Site. Hanford will likely offer the committee the greatest challenge because it is the oldest site, has many tanks that have leaked, and has the most complicated wastes because of the various management practices and several chemical processes that generated the wastes, including the earliest processing technologies. The committee may also extend the comments on the Savannah River Site found in this report as additional information on this site becomes available during the period of this study.