High-Level Waste Tanks at Hanford

The Hanford facility, built in the early 1940s in south central Washington state, was one of two centers of plutonium production for the US nuclear weapons program (the other was the Savannah River Site in South Carolina. Nine plutonium production reactors and five reprocessing plants that chemically separated plutonium from uranium and fission products were built at Hanford between 1943 and 1963. All reactors and reprocessing operations were shut by the late 1980s, though there have been periodic proposals to revive certain operations there, such as tritium production.

Hanford's five reprocessing facilities resulted in massive quantities of high-level liquid waste containing fission products (such as technetium-99, cesium-137, and strontium-90) and residuals of plutonium, uranium, and other heavy radioactive elements. The scale and complexity of Hanford wastes has made it the most difficult remediation problem in the United States. Approximately 54 million gallons (206,000 cubic meters) of high-level waste containing roughly 200 million curies of radioactivity are stored in 177 tanks at Hanford. (149 of these are single-shelled tanks, 28 are newer double-shelled tanks.) This represents 60% of total high-level waste in the United States by volume (the Savannah River tanks contain the largest amount of radioactivity, with about two-thirds of the total).

About 67 of the single-shelled tanks at Hanford have leaked or are suspected to have leaked. The volumes and radioactivity contents of these leaks are still the subject of considerable uncertainty. Official data have been published from time to time, with estimates of both volume and radioactivity generally increasing as new information comes to light (see tables 1 and 2).

Contamination of the Vadose Zone

The soil column above the water table around the tanks and below them, known as the vadose zone, has been contaminated by these leaks. Other dumping has also contaminated the Hanford vadose zone. For instance, large volumes of radioactively contaminated liquids were discharged into the soil and into "cribs" (trenches) built for the purpose. The highly contaminated vadose zone poses a severe risk to the most important surface water resource in the northwest, the Columbia River, which runs through the Hanford reservation. A failure to remediate the vadose zone and to empty the tanks of their radioactive waste would present a continuing threat to the region and its people and economy that could have unforeseeable negative consequences. DOE is moving some of waste from single shell tanks into double shell tanks to reduce the risk of leaks.

Recent data show that contamination from leaking tanks appears to be worse than previously thought. In August 1998, DOE released a report that examined leaks in the so-called "SX tank farm," concentrating on 5 tanks: 4 that have leaked and one that is believed not to have leaked.¹ The report estimates that 413,000 gallons of liquid contaminated with cesium-137 (a radionuclide with a half-life of about 30 years) have leaked from the four tanks, with a radioactivity level of 1 million curies (upper-bound estimate). The report gives a lower-bound estimate of about half this amount.

The report contains no analysis of the sensitivity of the results to variations in its assumptions about key parameters and notes that there is a great deal of uncertainty, but the new estimates of the volumes of waste that leaked are much higher than earlier ones. The radioactivity estimates are also higher. The previous estimate of the amount of cesium-137 in all the contaminated liquid that has leaked from all tanks was approximately 1 million curies. Table 2 shows various estimates of volumes of liquids that have leaked from these four tanks.

Efforts to establish a scientifically-sound approach to contamination of the vadose zone, begun recently by Undersecretary of Energy Ernest Moniz, must continue to receive high priority and attention. A thorough reconsideration of tank waste retrieval and tank decommissioning is also needed, since current plans appear to rely on groundwater models that have been invalidated by recent investigations and disclosure of data regarding radionuclide migration and leaks.

Tank Remediation

All of the leaking tanks at Hanford are "single-shell" tanks - i.e. tanks that do not have a second complete steel containment vessel enveloping the inner tank (see diagram, page 19). In total, the 149 single shell tanks (all beyond their design lives of 25 years) contain roughly 5,700,000 gallons of pumpable liquid. An important part of DOE's tank management involves pumping liquids from the single shell tanks into double shell tanks in order to prevent further leaks.

The process faces challenges, however. Liquids are present in the tanks as supernatant and interstitial liquid. Supernatant occurs on top of the sludge and saltcake (waste that has crystallized into chemical salts) in the tanks. Supernatant can be somewhat straightforwardly pumped from the tanks. But interstitial liquid occurs in the pore spaces of the saltcake and sludge and is more difficult to pump. In fact a considerable amount of liquid might remain in the pores even after extensive pumping. Therefore, it is difficult to ensure against leaks until the tanks are completely emptied.

The DOE has adopted the misleading practice of declaring a tank "interim stabilized" even if it still contains up to 50,000 gallons of interstitial liquid. Further, the DOE has no chemical or radiological criteria for declaring the tanks to be "interim stabilized." Since these tanks also contain flammable and/or explosive materials, and since the risk of fires depends on the amount of water present in the tanks, the pumping of liquids out of the single shell tanks (which include both water as well as other liquids) changes the risks both in single shell and double shell tanks. Hence, a declaration that a tank is "interim stabilized" should involve careful consideration of chemical and radiological criteria.

Although removing the liquids from single shell tanks is desirable in order to prevent further leaks, it also creates new concerns such as increasing the temperature in the tanks being emptied and changing the chemistry of the double shell tanks into which the liquids are being pumped. There is also a concern that the process of pumping out liquids may initiate new corrosion in the single shell tanks. As liquids are pumped, new parts of the inner wall of the tank are exposed at the point where the liquid and air meet (the "liquid-air interface"). Electrochemical phenomena that are not yet well understood could cause rapid corrosion at this interface.

Long-term management of tank wastes

In addition to the short-term goal of preventing leaks, it will be necessary in the long term to remove the waste from the tanks and put it into a form that will pose the lowest threat to the environment. DOE's current plan is to remove 99% of the waste volume from the tanks (and possibly more); separate the retrieved waste into high- and low-level waste streams; vitrify (turn into glass) both waste streams, disposing of the high-level waste in a geologic repository and the low-level waste on site. This plan has a number of problems, including that it will greatly increase the volume of highly radioactive waste dumped on the site.

Another problem is that the vitrification program is proceeding without sufficient technical preparation and without a proper back-up plan in case of failure. The DOE awarded a $6.9 billion "privatized" contract to British Nuclear Fuels, Limited (BNFL, a British government owned corporation) to vitrify waste in about 10 percent of the volume of Hanford tank wastes. The contract raises serious questions. First, the technology proposed by BNFL has not been adequately tested on Hanford's unique waste types. Second, construction of the vitrification plant would proceed when overall design work on the facility is less than 50% complete. If the technology fails, US taxpayers will pick up BNFL's costs.

The contract with BNFL also raises safety questions. Safety documents submitted by BNFL for the Hanford contract were described by DOE regulators as "poorly done."² In addition, BNFL's record in its home country, where it is covered by the British Official Secrets Act, leaves much to be desired. The DOE has not used the leverage of contracts with BNFL's US subsidiary to raise the issue of making the records of BNFL's British operations public. We believe making these records public is relevant to assessing how it will perform in its US operations.

Because this plan would involve disposal of the vitrified "low-level" waste at Hanford, DOE envisions that waste going to a deep repository would be reduced. DOE has failed to account for the cost of increased local disposal at appropriate open market equivalent prices. Moreover, the so-called "low-level" waste designated for on-site disposal would, in other countries such as Britain or France, be classified as "intermediate level waste" and be designated for deep geologic disposal.

Finally, DOE does not appear to be planning for the decommissioning of the tanks themselves. Rather, the plan appears to call for pouring cement into the tanks after they have been pumped out, even though that process may leave up to one percent of the volume of the highly radioactive waste in the tanks. The radioactivity in this waste could, in many tanks, present a serious long-term environmental hazard. If the waste leaks from the tanks, cementation of the tanks will have created a huge new problem that could greatly complicate any future attempts to remediate the vadose zone.

Of all Cold War wastes in the United States, those at Hanford are the most varied and the problems they pose are the most intractable. According to recent estimates, removal and treatment of Hanford tank wastes will cost about $15 billion. Even this huge amount overlooks several costs, such as those required to decommission the tanks themselves, deal with the contaminated soil around the tanks due to direct discharges and leaks, and remediate the contaminated groundwater. It also does not account for the cost for possible vitrification technology failures, discussed above.

Recommendations

The Hanford tank program needs to be thoroughly revamped. It should shift from the present arbitrary goals to ones that are better suited to environmental protection, and to short- and long-term waste management and disposal. For example, for the purposes of interim waste stabilization, DOE should examine calcining, an approach to solidification of waste that would involve heating the wastes and turning them into a powder form. Calcining would result in a relatively stable waste form and greatly reduced waste volumes and is therefore likely to be more compatible with either vitrification or immobilization in ceramics. Calcine can be stored without the same kind of serious short- and medium-term risks to the environment associated with the current form of tank waste. Despite these potential advantages, the DOE and its contractors have not carefully examined the option of using calcining as an interim method. Rather, they have dismissed it by noting that calcining would not produce a waste form suitable for repository disposal, a fact not in dispute.

Other IEER recommendations are that the DOE examine the following elements more carefully than it has done so far:³

• Adopt a goal to process all high-level waste tank contents for management as high-level waste;

• Revamp groundwater models to reflect actual data on vadose zone contamination;

• Initiate two parallel programs for solidification of high-level waste: 1) develop methods for calcining high-level waste while researching ceramic and glass immobilization for the calcine, and 2) pursue pretreatment and specific glass-making approaches that would not require calcining. Additional IEER recommendations for waste treatment at Hanford address determining the extent of existing contamination in light of decommissioning and decontamination plans (see report).