Immobilization of Excess Weapon Plutonium: A Better Alternative to Glass

The United States plans to immobilize several metric tons of excess weapons plutonium in a solid matrix. The selected material must achieve the short-term goal of deterring proliferation through theft or host-nation reuse, and the long-term goal of preventing plutonium exposures over geologic time. The Department of Energy, after internal review, has recently decided on Synroc, a crystalline ceramic, to immobilize the plutonium. This paper presents an independent technical comparison of glass versus ceramic immobilization options, and reaches similar conclusions to those of the Department of Energy. On a technical basis, Synroc performs better than glass in a number of areas. It is more proliferation resistant than glass due to the more complicated and less well-known extraction process that would be required to separate the plutonium. Synroc is more chemically durable than borosilicate glass and can dissolve more depleted uranium than glasses to address future criticality problems. Now that the Department of Energy has selected Synroc as the waste form of choice for plutonium disposition, it should also be reconsidered for immobilization of high-level nuclear waste.

Immobilization of Excess Weapon Plutonium: A Better Alternative to Glass

Allison Macfarlane

INTRODUCTION

Both Russia and the United States are faced with decisions on how to dispose of plutonium and highly enriched uranium recovered from dismantled warheads, and from various nuclear weapons facilities. In the U.S., the Department of Energy (DOE) estimates that 50 or more metric tonnes (MT) of weapons plutonium and hundreds of metric tonnes of highly-enriched uranium (HEU) will be considered "excess." Disposition of these materials is essential for national and international security reasons. The disposition of excess HEU is a relatively straightforward process in which it is mixed with natural uranium to dilute it to the low-enriched composition used in commercial nuclear reactor fuel. In contrast, the disposition of plutonium is not so simple; dilution with other plutonium isotopes will not render it "safe" because all isotopes of plutonium can be used to make nuclear weapons. However, the National Academy of Sciences' Committee on International Security and Arms Control1 advanced the "Spent Fuel Standard" criterion to provide a guidepost for plutonium disposition. The Spent Fuel Standard requires that the weapons plutonium be converted to a form as inaccessible as plutonium in spent fuel from commercial nuclear reactors.2

In addressing excess weapons plutonium disposition in the United States, the Department of Energy (DOE) and the Clinton Administration have adopted a dual-track approach that will give the DOE the option both (1) to mix plutonium oxide with uranium oxide to form MOX fuel, which will be burned in selected light-water reactors, and (2) to immobilize plutonium in solid form such as glass or ceramic.3 It is the latter option that forms the focus of this paper. Despite the dual-track approach, it is certain that at least 17 MT of the excess plutonium will be immobilized rather than converted to MOM This is a substantial part of the plutonium metals, scraps, and residues located at 8 major sites in the United States including Pantex in Texas, Rocky Flats in Colorado, and Hanford in Washington (Figure 1)4 which are in forms that would be prohibitively expensive to purify sufficiently for use as MOX fuel.5

Immobilization of excess weapons plutonium will almost certainly use a can-in-canister method, where 20 small stainless steel cans containing plutonium incorporated in ceramic or glass will be loaded onto a frame inside a large (3 m long by 0.6 m diameter) stainless steel canister (Figure 2).6 Each can will contain about 2.56 kg of plutonium immobilized in the waste form, for a total of 51.2 kg Pu per canister.7 After the small cans are loaded onto the frame, borosilicate glass containing defense high-level nuclear waste will be poured into the large (1,800 kg) canister, and a lid will be welded onto the canister. The high-level waste glass contained in the outer canister will provide a gamma radiation barrier to deter theft of the plutonium.

The canister design, employing borosilicate glass, will be common to whichever waste form-glass or ceramic-is used for the can; and it is the choice of this waste form that the remainder of the paper will examine.

Because plutonium is extremely toxic, a grave proliferation risk, and has a half-life of 24,100 years, it is essential that the waste form into which it will be incorporated be able to both (1) prevent reuse in nuclear weapons by both host nations and terrorists and (2) safely contain it for a very long time to prevent exposures to humans and the environment. Although the United States along with many other major nuclear countries has favored the use of glass for high— level nuclear waste immobilization, plutonium, a fissile material, requires additional considerations, and there now appear to be a number of waste form options that have significant advantages over glass. This paper will focus on two types of waste form, crystalline ceramics and glass, and compare their properties and production technologies.

Figure 1: Map of locations of excess weapons Pu in the United States.

It is widely assumed that the immobilized waste is eventually destined for a geologic repository. For this reason, it is important to try to understand the repository environment over time-scales greater than a few thousand years. However, it is actually easier to model the behavior of some waste forms over a geologic time-scale than it is to model the hydrology or seismicity associated with the repository environment. Past repository performance assessments have concluded that plutonium would not be a major contributor to environmental radiation doses on the basis of the low solubility of plutonium put in groundwater (on the order of 10-8 gm/cm3).8 Recent findings on plutonium mobility at the Nevada Test Site, however, suggest just the opposite, that rela-tively rapid transport of plutonium in the repository environment may occur from colloidal transport of species.9 In light of the uncertainty associated with plutonium transport in geologic media coupled with large uncertainties in performance assessments of repositories (due to uncertainties in future volcanic activity, seismicity, flow in the unsaturated zone, climate change, etc.) it would be prudent to choose a waste form that ensures containment of radionuclides, regardless of repository performance. This conclusion was reached by the National Research Council's Committee on Vitrification of Radioactive Wastes, which stated that "sensitivity analyses used to evaluate waste form perfor-mance should emphasize the material properties of the waste form, not the total system performance."10 This is the approach taken in this paper.

Figure 2: Schematic of the can-in-canister design. Twenty cans of plutonium-containing waste form are attached to a frame and inserted into a larger canister. The canister is filled with high-level radioactive waste glass to provide a radiation barrier for nonproliferation purposes.

Nevertheless, the waste form should to some degree be appropriate for the particular type of geologic repository. Yucca Mountain, Nevada, the location of the planned high-level waste geologic repository, is composed of tuff, a fine-grained silicic rock solidified from volcanic ash flows. It is dominated by SiO2, Al2O3, K2O, NaO, and CaO, as is crystalline rock, such as granite or gneiss, under consideration in Europe as repository media. This contrasts with bedded salt, the repository lithology at the Waste Isolation Pilot Plant in Carlsbad, New Mexico. Because of the different lithologies, the groundwater compositions, pH, and redox conditions will vary in these different geologic environments. The selected waste form must be able to perform well under the conditions that will exist.

AN INTRODUCTION TO THE WASTE FORMS

Over the years, researchers have considered a number of materials to encapsulate high-level nuclear waste. These materials include, but are not limited to, glasses of various compositions, including silicate glasses11 and phosphate glasses;12 ceramics of a wide variety of formulations;13 glass-ceramic compounds;14 cements;15 coated particles or particles in metal matrices;16 and pyroprocessed metals.17 Based on present familiarity with production technology, the two best candidate materials for high-level nuclear waste and plutonium disposition are glasses and crystalline ceramics.18

Glass
Glass is a noncrystalline solid in which a wide range of waste impurities or material can be dissolved.19, 20 Glass may be relatively susceptible to damage from radioactive decay, although this has yet to be proven experimentally.21 As discussed in the Appendix, a number of countries have developed successful industrial-scale vitrification technologies to solidify their HLW.22 For these reasons, glass has received the most attention as a potential waste form for the immobilization of plutonium. On the other hand, glass is thermodynamically unstable and over geologic time may devitrify (crystallize), especially at the elevated temperatures (100o-300o C) expected to be encountered in a geologic repository.23

Table 1: Compositions of glasses for plutonium immobilization in weight percent.

a. Standard Defense Waste Processing Facility borosilicate glass composition.24
b. Lanthanum borosilicate glass based on loffler glass composition, developed at Lawrence Livermore National Lab,25

Plutonium has already been added to this composition, unlike the other LaBS gloss.

c. Lonthanum borosilicate glass based on loffler glass composition, developed at the Savannah River Site.26

Borosilicate glass, as opposed to other compositions such as phosphate glass, is preferred for high-level nuclear waste. It has the advantage of being more durable than many other glass compositions and can be produced at lower temperatures than other glasses.27 Table 1 shows three glass compositions that may be suitable for excess weapons plutonium material. In the standard glass composition being used at the Savannah River Defense Waste Processing Facility (DWPF), B and Li are present to improve the properties of the glass, such as lowering the viscosity of the melt. The glass composition for plutonium disposition is a lanthanide borosilicate (LaBS) glass because of the higher solubility of plutonium and greater chemical durability. Overall, glass can accept a wider range of impurities than ceramic, but with limits. For example, LaBS glass can accept only a limited amount of uranium, as will be discussed in a later section of this paper. Borosilicate glasses are commonly processed between 1,100o-1,200o C whereas LaBS will be processed at 1,475ïC.28 For the DWPF borosilicate glass composition of Table 1, waste loadings for plutonium are on the order of 2-4 wt percent.29 One study suggests that higher waste loadings in borosilicate glass, on the order of 7 wt percent, would be possible.30

Two formulations of lanthanide borosilicate (LaBS) glasses are shown in Table 1. Gadolinium and Hafnium in the LaBS glass are added as neutron absorbers. Lawrence Livermore National Laboratory developed the LaBS 1 composition, in which plutonium is already part of the composition. The LaBS 2 glass, developed by the Savannah River Site, will be used as a base composition from which the final composition will be developed. The measured solubility of plutonium in the LaBS glass at 1475oC was observed to be greater than 10 wt percent in stirred crucible experiments, although a maximum of 8.5 wt percent Pu was observed in an unstirred crucible melt of a 1 kilogram monolith of LaBS glass.31

Natural analogs of waste form material provide the only data available data on long-term performance of these materials under geological conditions. Many geologic examples of glass exist in the natural world. These glasses fall into two compositional categories: those highly enriched in silica, such as rhyolitic glasses and tektites (70-75 wt percent Si02) and those less enriched in silica, such as basalt glasses (45-50 wt percent Si02).32 Tektites are impact glasses that form under extreme conditions of temperature and pressure, whereas rhyolite and basalt glasses form during extrusive volcanic events (low pressure, high temperature). Natural volcanic glasses are generally much younger than 40 million years of age.33 A few rare glasses, lunar glasses, are much older, on the order of 108 years, but have probably survived for so long because of their lack of contact with water in the lunar environment.34 Comparison of the silica content of these glasses with the LaBS glasses (29-35 wt percent SiO2) in Table 1 indicates that neither the rhyolites or the basalts would be reasonable natural analogs for the LaBS glass composition. For HLW on the other hand, basalt glasses would be justifiable analogs to the DWPF borosilicate glass in Table 1. Few natural glasses contain alpha-emitters in large enough quantities to provide a reasonable estimate of the effect of radioactivity on glass over geological time.

Table 2: Compositions of ceramic waste forms for plutonium immobilization.

a. Classic synroc-C composition35
b. Most recent formulation of synroc compostion for plutonium disposition at Lawrence Livermore National Laborotory.36
c. May or may not be present with zirconolite.
d. For synroc-C, these are Ti oxides (rutile) and Co-AI titanates and alloys,37 for LLNL synroc, this is Hfo2 Or (U,PU)0238

Ceramics
Ceramics, in contrast to glass, are crystalline materials in which radionuclides from nuclear waste are accepted into the crystal structure by substituting for the components that constitute the phase. Consequently, it is usually possible to predict where an element will go in the crystal structure, based on its ionic radius and charge. A number of different compositions of ceramics may be appropriate for the immobilization of nuclear waste, such as zircon, monazite, apatite, baddelyite, cubic zirconia, or Synroc (Table 2).39 To geologists, crystalline ceramic materials are simply minerals or mineral assemblages, the building blocks of rocks. Many natural analogs exist for the ceramics suggested above.

For plutonium immobilization, Synroc, zircon, monazite, baddelyite, or cubic zirconia may be appropriate waste forms. For simplicity of discussion in this paper, I will focus on three of the most promising possibilities, Synroc, zircon, and monazite. Zircon and monazite are single-phase species that in nature contain significant amounts of actinides (zircon contains U and monazite, U and Th). In fact, both minerals are common in crystalline rocks and are used by geochronologists for the determination of age of crystallization or metamorphism with the U-Pb method of dating. Zircon from the Canadian shield was used to determine the age of one of the oldest known rocks on earth, a 4.02 billion-year-old gneiss.40 The oldest known monazite is over 2 billion years of age.41 These phases are extremely durable, resistant to corrosion and radiation damage, and consequently have survived for extremely long periods of time.

The best known multi-phase ceramic waste form, Synroc, is composed of some combination of the following minerals: zirconolite, pyrochlore, hollandite, perovskite, rutile, and minor oxides and alloys.42 Synroc-C included hol-landite, zirconolite, perovskite and alloys, and was designed to accommodate high-level waste from the reprocessing of spent nuclear fuel from commercial power reactors (Table 2).43 Hollandite hosts fission products such as Cs, Ba, Rb, K, and Cr.44 The tetravalent actinides, Th, U, Pu, and Zr are immobilized in zirconolite or pyrochlore, whereas perovskite accommodates Sr, Na, trivalent actinides and rare earth elements. The alloy accepts Tc, Mo, Ru, Pd, S, and Te.45

Although Synroc itself does not exist in nature, all of the major phases in it do. Zirconolite is found in igneous rocks such as carbonatites and ultramafic assemblages. The oldest known zirconolite, over 2.5 billion years old, is from a layered mafic complex in Australia.46 Pyrochlore, known as a source of rare elements such as Nb, Ta, and W, is found in carbonatites and both nepheline syenites and granite pegmatites.47 Pyrochlore is known to range in age from 16 million years to at least 1.4 billion years old.48 Perovskite is usually found in rocks with low SiO2 content and thus is a common mineral in the upper mantle of the earth. Hollandite is found in rare volcanic rocks with compositions high in K2O, BaO, and Ti02 and low in SiO2.49 Rutile is a very common accessory mineral found in a wide variety of rock types.

Table 2 gives the compositions of these waste forms. It includes two formu-lations for Synroc-the standard Synroc-C formulation, developed by Ringwood50 and the most recent compositions used in trials at Lawrence Liv-ermore National Laboratory (LLNL).51 One of the LLNL formulations is predominantly pyrochlore, of which zirconolite is a derivative structure, and could accommodate a high percentage of plutonium. Lawrence Livermore National Laboratory has recently done experiments on the incorporation of impurities into Synroc. For the expected range of impurities and impurity loading based on the waste stream, Synroc accommodates them by forming the phases pyrochlore, zirconolite, brannerite ((U, Pu) Ti206), rutile/silicate, and an actinide oxide in varying amounts.52 Pyrochlore has the additional benefit of accepting large quantities of U into its structure (U is one of the major elements in pyrochlore), which will offset two problems: that of high quantities of U in the waste stream and the possible need to purposefully include depleted U to decrease the likelihood of a criticality event from U-235 after Pu-239 decays (further detail on this topic is covered later in this paper).

The selection criteria for a waste form to immobilize excess weapons plutonium must address both short-term and long-term goals: to secure plutonium from reuse by both terrorists and host nations and to prevent doses of plutonium to humans and the environment, respectively. In the short term, the waste form should (1) be resistant to recovery of plutonium and (2) have a mature production technology and low production costs and timing. In the long-term the waste form must have (1) good chemical durability over geologic time, (2) an ability to accommodate radiation damage from radioactive waste, and (3) an ability to avoid criticality.

The Appendix presents a systematic comparison of glass and ceramics in terms of these criteria.

CERAMIC AND GLASS: A COMPARISON OF UNCERTAINTIES AND ADVANTAGES

Some of the questions associated with Synroc and LaBS glass issues may be resolved in the short term, whereas other issues require intensive, long-term work. Table 3 displays a comparison of uncertainties concerning LaBS glass and Synroc. Uncertainties associated with glass fall into the categories of glass production, chemical durability, and radiation damage. In glass production the largest uncertainties are related to the reliability and safety of the high-temperature melting process behavior of the glass during the first and second glass pours, such as the effects of glass fracturing on chemical durability, and the significance Of PuO2 crystallization. Experiments on the effects of water on microfractures in glass and others on the behavior of crystalline PuO2 with respect to chemical durability, radiation damage, and proliferation resistance will most likely require more than a few months to complete, if any have been planned by the DOE.

In terms of chemical durability, the effect of colloid behavior on the transport of actinides, and in particular, plutonium, is another uncertainty associated with LaBS glass, which will not soon be resolved. LaBS glass is a new composition about which there is scant available data. As a result, we know little about the type and conditions of formation of colloids and less about their ability to bind up plutonium and transport it. Many leaching experiments require long reaction times (years). Another long-standing and unresolved issue is that of radiation damage to glass, specifically to the LaBS glass composition. I would argue that our understanding of radiation-damaged glass is in the early stages. We are still learning which questions are appropriate to ask, and we will learn more only after considerable research on the topic. Furthermore, no natural analogs exist to provide an understanding of radiation damage in glass over geologic time.

In ceramics a number of uncertainties also await clarification. As with glass, the effect of radiation damage on the chemical durability of ceramics remains an unresolved question. With ceramics we have the advantage of using natural analogs as an indication of the effects of radiation damage. Although most radiation-damaged minerals remain chemically durable with regard to actinide retention, the underlying question of radiation effects from plutonium over geologic time may not be resolvable. Like glass, the role of colloids in leaching of ceramics requires more research. Another uncertainty associated with ceramic production involves the effects of density on chemical durability for current Synroc formulations. If lower density Synroc is found to be deficient, then variations in production methodology should be investigated to address the density problem. It will take considerable time to perform adequate leach tests to assess the affects of low density. The ability of Synroc to produce plutonium-binding colloids during alteration also requires attention over the long-term.

Table 3: Uncertainties associated with glass and Synroc.

Ceramics and glass appear to be basically equivalent in terms of maturity of production technology, timing of waste form production, and in cost of pro-duction. However, with respect to other criteria, each waste form appears to have certain advantages. These are summarized in Table 4.

Table 4: Advantages of glass and Synroc.

Glass has advantages over ceramic in its theoretical ability to accept a wide range of impurities into its structure and its potential ability to withstand radiation damage. Ceramics and Synroc in particular, on the other hand, are proven to have corrosion rates at least one or more orders of magnitude lower than glass, and thus should better contain the waste over the geological time frames necessary for plutonium disposition. In addition, Synroc is more stable thermally than glass, and, unlike glass, its chemical durability will not be affected by high temperatures that may be encountered in the geologic repository. Furthermore, Synroc is able to withstand the temperatures of reheating that the waste form will experience during the second pour of HLW glass with only minor effects. Synroc can easily accommodate more depleted U than glass to address the problem of decay of Pu-239 into mobile U-235. The dilution of the enriched U by depleted U will help prevent future criticality events that might be caused by U-235 in a geologic repository. Ceramics in general are not soluble in nitric acid, unlike glass, and accordingly the PUREX process cannot be easily adapted for plutonium extraction. The failure modes of Synroc production are more benign than those of glass production.53 Worker safety should be an important consideration for a large-scale effort for excess weapon plutonium disposition.

THE POLICY CHOICE

The criteria that form the basis for the waste form decision should be proliferation resistance, production technology, chemical durability, and criticality safety. In the short term, proliferation resistance is the most important selection criterion. Ceramics (in particular, Synroc) are more resistant to extraction of plutonium than glass. This basic difference sets ceramics apart from glass. Because extraction of plutonium from ceramics would require major modifications to presently-existing separation facilities and construction of new ones, such activity would be easier to detect through safeguards than extraction from glass.

Among the selection criteria that apply to the long-term, chemical durability and criticality safety are the most significant. Considering the large uncertainties in predicting future geologic events and the state of a repository environment, we should rely on the material most likely to resist alteration and corrosion and impede the release of actinides into the environment, which is ceramic.

Whereas criticality safety was not a criterion in the selection of a waste form for HLW immobilization, it should be for plutonium disposition. To prevent criticality events due to Pu-239's mobile daughter-product, U-235, depleted U should be added to the waste form to dilute U-235. Only ceramic can adequately support the addition of large amounts of actinides to its composition.

In September 1997, the Materials Disposition office at the DOE announced its decision to use Synroc to immobilize excess weapons plutonium.54 The decision was based on five criteria, and in four of these areas the DOE judged that ceramic held advantages over glass.55 The criteria and findings were similar to those of this paper. In evaluating waste form performance, criterion (1) focuses on the repository environment. DOE found that ceramics would be more durable than glass, and a DOE panel found that ceramic would provide better criticality assurance because it can incorporate more U-238 than can glass.56 Criterion (2) was environment, health, and safety. DOE disclosed that workers handling the glass product would be exposed to eight times the neutron dose rate of the ceramic product due to alpha-n reactions from boron in the glass.57 Criterion (3) was costs, and it was noted that increased worker protection against radiation from handling the LaBS glass would result in increased costs. Also, costs for LaBS glass would be higher than for ceramic as a result of the higher loading of plutonium in ceramic. Overall, this would result in fewer cans of immobilized plutonium requiring fewer canisters of high-level waste glass. For the final criterion (4), nonproliferation, DOE found that ceramic provides an advantage over glass due to its greater resistance to plutonium extraction. DOE noted that the plutonium extraction process from ceramic is more complicated and less well developed than that from glass. In terms of detection of theft or diversion of plutonium during waste form processing, ceramic provides the additional advantage of easier non-destructive assay Alpha-n reactions from B in glass interfere with the verification of plutonium concentration.

Although DOE is currently advancing ceramic as the waste form of choice for plutonium immobilization, it does so by issuing tepid statements on the advantages of ceramic over glass. Overall, the DOE emphasized that the advantages of ceramic over glass were small. Even though there was a shortage of data on the two waste forms to make an adequate comparison, DOE moved the waste form decision up a year, and clearly not all necessary experiments could be completed in such a shortened time frame.

CONCLUSION

Recently, both LaBS glass and Synroc ceramic were competing waste forms for the can-in-canister option to immobilize excess weapons plutonium. On the basis of the technical information presented in this paper, ceramics are the preferable waste form for plutonium immobilization. Ceramics perform better than glass in terms of their chemical durability, their thermal stability, their proliferation resistance, their ability to withstand heating that they will encounter during the pouring of the canister HLW glass, and their ability to dilute U-235 formed from the transmutation of Pu-239. Synroc ceramic is equivalent to LaBS glass in terms of maturity of production technology and timing of production according to DOE estimates.

This comparison can be extended to the immobilization of HLW. Although this material has different properties than plutonium (for example, higher beta and gamma emission and volatilization problems), it is time to reconsider ceramics as a waste form for HLW. For long-term disposition, the main concern for HLW, ceramics are superior to glass due to their greater chemical durability. We should begin to explore in earnest other promising phases such as zircon, monazite, and baddelyite and develop new Synroc assemblages. Nuclear waste and fissile materials are presently abundant and it is our responsibility to deal with them in the best manner available.

APPENDIX: SELECTION CRITERIA FOR THE WASTE FORM

RECOVERABILITY

Unfortunately, no completely irreversible disposition methods exist for plutonium. Nonetheless, the form and design used should be the most irreversible possible. If it is more difficult for the host nation to extract plutonium from the immobilized form than spent fuel, both in terms of cost and institutional capability, then the waste form is successful. In comparison to host nations, it is assumed that terrorists would want a much smaller quantity of plutonium. However, even one canister of immobilized plutonium would contain enough material for 12 weapons (at 4 kg Pu per bomb).

In the case of both glass and ceramic, the can-in-canister design will discourage terrorist attack by the sheer mass of the canisters (each will weigh over a ton) and the radiation barrier of 200-500 R/hr at 1 m from the surface of the canister 30 years after fabrication of radioactive HLW glass.58 If terrorists were able to steal these canisters, they may somehow be able to break the cans containing plutonium out of the radioactive canisters, but this is much more likely accomplished by a host nation, which would have the facilities, institutional networks, and funding to organize such a process. If the design is successful and the cans are an integral part of the canisters, then both host nations and terrorists would have to handle the material remotely, and would have to dissolve the entire canister to extract plutonium. This process would be easier for host nations, who have the institutional support, but it would prove expensive for both parties.

Two recent DOE reports have criticized the can-in-canister design as not adequately meeting the spent fuel standard .59 The Red Team Report60 claims that the cans are mechanically separable from the HLW glass relatively rapidly and with simple equipment. Once the radiation barrier is gone, the plutonium-containing cans could easily be removed and handled because shielding would no longer be required. The DOE is now beginning tests to ascertain how separable the cans are from the glass matrix.61 If the cans are determined to be easily separable, a new design would be required in which the cans are made an intrinsic part of the HLW glass. This could be accomplished by making the cans fracture with the HLW glass if the canister is breached or by using small pellets of plutonium-containing ceramic or glass. 62

A recent study of the proliferation resistance of borosilicate glass suggested that a critical mass of Pu (4.7 kg) could be recovered from only 613 kg of borosilicate glass with a waste loading of 2 wt percent Pu using simple benchtop methods (a recovery rate of 27 percent).63 For comparison, one canister of glass is expected to weigh approximately 1,800 kg,64 suggesting that for homogeneously mixed plutonium at 10 wt percent, it could produce over 70 kg plutonium. To extract plutonium, the glass would first have to be crushed, ground, and dissolved in nitric acid, and then the PUREX process could be applied to separate plutonium.65 To this end, the United States could augment already-existing PUREX facilities at Hanford, the Savannah River Site, and the Idaho National Engineering Laboratory with new large-scale crushing, grinding, and dissolving facilities. Consequently, it would be about as easy to extract plutonium from a glass waste form as from spent fuel.

More importantly, LaBS glasses were developed to serve as a temporary storage form for the actinides americium and curium, which have commercial value. Recent research on the dissolution of LaBS and Sr-Al-borosilicate glasses to recover americium and curium was conducted by Savannah River Site scientists. Erbium was used as an analog for Am and Cm in the experiments in which the glasses dissolved completely after 2 hours in nitric acid at 110ïC.66 Workers were able to recover 100 percent of the lathanides from this process. Thus, from the point of view of the host nation, LaBS glass offers little resistance to extraction of plutonium, certainly the same or less than that of spent fuel.

The proliferation resistance of ceramics is not as well documented as that of glass, although one DOE study suggests that ceramics would be more difficult to grind and dissolve.67 Ebbinghaus and others68 claim that titanate— based ceramics, like those used in Synroc, are not soluble in nitric and hydrofluoric acids, those used in the PUREX process. The acids in which Synroc is soluble actually interfere with the PUREX process, and consequently the process cannot be easily adapted for use to extract plutonium.69 The difficulty in dissolving ceramic materials is illustrated by the techniques used to put these materials into solution for industrial and experimental purposes. Peterson and others70 investigated the use of perovskite, a component of Synroc-C, as a source of titanium. They achieved over 90 percent recovery of titanium for dissolution in acid concentrations above 70 percent H2SO4 (with best results for solutions above 90 percent H2SO4), at temperatures above 200oC for 35-mesh size particles and above 150oC for <100-mesh-size particles.71 The time required for dissolution ranged from a few minutes to 6 hours, depending on the above parameters. Similarly, zircon and monazite require complete dissolution for the U-Pb method of geologic age-dating. The experimental procedure for dissolving these materials follows a well established procedure of combining the minerals with 6N HIP and a small amount of HN03 in a teflon "bomb" at 240oC for 2 days.72 The teflon "bomb," a screw-top teflon bottle, creates a pressurized environment when heated.

Based on the above data, the only proliferation resistance offered by LaBS glass is that of the canister design, which will be common to both waste forms. Ceramics, on the other hand, offer the additional resistance of having no established industrial-scale processes or facilities to dissolve and extract plutonium. Acids such as HF and H2SO4 are necessary for plutonium extraction, but these acids are not used in the standard PUREX process. As a result, a host nation that was determined to extract plutonium from its ceramic waste form would need to build new facilities and develop new technologies to extract large quantities of plutonium. In comparison to extraction of pluto-nium from spent fuel, then, it is more difficult to extract plutonium from ceramic, but as easy or easier to extract plutonium from LaBS glass.

Production Technologies

There is already much experience in the production of glass to contain HLW (see Figure 3). Initially, two main vitrification processes were developed, one by France (AVM) and one by Germany.73 Glass production in France began in 1969 with the PIVER (Pilote Verre) plant, which operated until 1972, producing a total of 12 MT of glass.74 In 1978, the Atelier de Vitrification de Marcoule (AVM) facility began producing HLW glass using a continuous vitrification process and by 1995 had produced 857.5 MT of glass.75 The feed solution is first dried and calcined and then mixed with glass frit in an induction-heated metal melter. The U.K. adopted the AVM process for its Waste Vitrification Plant at Sellafield.76

The PAMELA facility in Belgium, begun in 1985, uses the German pro-cess. The material is vitrified in a single-step in a ceramic melter, where glass frit and HLW are loaded into the melter and drying, calcining, and melting occur in sequence.77 China, still in the R &D stage, plans to adopt the German process.78 Russia has vitrified HLW in phosphate glasses at their Mayak facility where operations began in 1987, operated for 1.5 years, and restarted in 1991, producing a total of 1,800 MT of glass by 1995.79 Russia is now developing a borosilicate glass to handle an increased radionuclide content.80 Japan is also in the process of developing a borosilicate glass composition for its newly-built vitrification facility at Tokai, although this facility may have been affected by recent shutdowns.81

The United States presently has two operating vitrification facilities, the Defense Waste Processing Facility (DWPF) at the Savannah River Site and the West Valley, New York, facility. The DWPF, which began processing radioactive waste from defense installations in April 1996, employs a complex procedure. The HLW must first be processed to remove mercury and organic materials prior to mixing with glass frit in a melter.82 The DWPF will produce 6,000 canisters of HLW glass, at 1,800 kg per canister, over the next 25 years.83 In contrast, the lifetime of the West Valley facility will be only 2.5 years, and the metal melter will produce approximately 300 canisters.84 Other vitrification facilities are in the R & D phase.

Figure 3:Map of vitrification technologies and accomplishments in various countries.

However, although HLW glass has been produced on an industrial scale for the past few decades, the technology to immobilize plutonium in glass requires special conditions, such as glove-box-size apparatus and significantly higher processing temperatures and is not as advanced as regular borosilicate glass technology.85 Although ceramics have not been used to immobilize HLW at an industrial scale in any country, a group at ANSTO (the Australian Nuclear Science and Technology Organization) have been producing Synroc using large-scale production technology at the Synroc Demonstration Plant (SDP) since 1987,86 and since this technology may be used as a benchmark to compare with glass production technology, such a comparison of production technologies for ceramics and glass suggests that they are at similar levels for the immobilization of plutonium. In fact, DOE estimates of production time schedules for the can-in-canister option for both ceramic and glass cans are exactly the same.87 Cost estimates for the ceramic and glass can-in-canister options are also identical.88 On the basis of cost and scheduling estimates, it appears that the DOE sees little difference in the difficulty of production for either glass or ceramic.

LaBS glass will be produced using melter technology substantially modified to handle fissile materials. Workers at the Savannah River Site are planning to use bottom-pour, inductively heated melters, with platinum-rhodium crucibles and platinum stir rods.89 To the LaBS glass base composition will be added 1.5 kg of 5-10 ìm PuO2 powder.90 The residence time in the melters is approximately 4 hours at a temperature of 1,475oC.91 The melters will be scaled for glove-box work, and the geometry of the melters will provide some prevention against criticality events.

A number of pitfalls face the planned glass production technology. First is the safety issue of failure modes. In a melter system failure, workers are dealing with a plutonium-rich corrosive liquid at temperatures of 1,475oC, which has the potential to be extremely dangerous. Secondly, devitrification and crack formation result from quenching the glass melt. Further devitrification and crack formation occur during the pouring of the canister HLW glass. Cracking, as discussed earlier, can lead to accelerated rates of leaching. In a number of experiments, the second glass pour resulted in the formation of PuO2 crystals in the LaBS glass.92 The presence of these crystals may not affect the long-term performance of the material, but we know little about the behavior Of PuO2 in a geologic environment. Not much information is available on the chemical durability Of PuO2 (will plutonium be leached out of the solid), radiation damage effects on PuO2, or the proliferation resistance of PuO2 (how easy is it to dissolve and extract plutonium). In addition, formation Of PuO2 crystals raises questions about the homogeneity of PuO2 in the glass. It is important for criticality and nonproliferation reasons to ensure that PuO2 is distributed homogeneously in the glass. Areas of concentrated PuO2 crystals may not receive adequate neutron absorption, leaving open the possibility of criticality. Glass may pose a problem in terms of materials counting and accountability in that the plutonium needs to be distributed homogeneously to detect it with low uncertainty. The Red Team Report suggests that this homogeneity in glass may be difficult to achieve.93

Ceramics can be produced by a number of methods such as cold pressing followed by sintering, hot pressing, and hot isostatic pressing. At ANSTO, the Synroc Demonstration Plant has been on-line since 1987 and has fabricated more than 6000 kg of Synroc with simulated HLW.94 The method they follow is to calcine the acid HLW and Synroc-C powders at 700oC for 1-2 hrs., add 2 wt percent Ti powder, load the mix into bellows, and hot press at 1,150o,1200oC at 14-21 MPa for 2 hrs.95 This method produces Synroc-C at 99 percent of its theoretical maximum density.96 In contrast the Department of Energy intends to use a cold-press-and-sinter method to produce Synroc, similar to the production methodology for MOX pellets for nuclear fuel. Hot pressing usually produces a denser material, but it is more expensive than cold pressing. At the Lawrence Livermore National Laboratory, Synroc is processed by first mixing 5 µm-size component powders and calcining them at 600oC for 1 hour, then 1-5 gm-size PUO2 powder is added and the mix is cold pressed at 15-20 MPa and then sintered at 1,350oC for 4 hours.97 The final product has greater than 90 percent of its theoretical maximum density.

Synroc production is not without its own set of pitfalls. One disadvantage it shares with glass production is that of the fine particle size of the PuO2 and other component powders. One to five microns is a respirable size and these powders have a tendency to disperse within the glove boxes used to handle them. Prior to the initiation of ceramic or glass production the DOE must deal satisfactorily with this issue. In addition, Synroc that is produced by cold pressing and sintering reaches only 90-95 percent of its theoretical maximum density. The most serious problem resulting from low density Synroc is its reduced chemical durability over the long term. Static leaching experiments on Synroc suggest that leach rates are lowest for densities of at least 98 percent; the leach rate for some species in Synroc increases by at least 2 orders of magnitude for a density decrease of 98 percent to 90 percent.98 At present, there is no information on the leach rates of plutonium as a function of density for the proposed composition of Synroc. As a result, it is important for the Department of Energy program to try to increase the density of the plutonium-containing Synroc.

The Lawrence Livermore National Laboratory has also done a number of experiments trying to produce zircon and monazite by the cold-press-and-sinter method, with only partial success. Zircon component powders that were cold pressed and sintered at 1,650oC for 1 hour resulted in theoretical densities of 65-75 percent.99 Monazite that was sintered at 1,350oC resulted in a theoretical density of 71 percent.100 In neither the zircon nor monazite experi-ments did all the material react to form the intended phases. Ultimately, zircon and monazite may be promising waste forms for HLW and plutonium in terms of chemical durability and ability to withstand radiation damage, but their production technology is too immature to immobilize excess weapon plutonium in a timely manner.

Chemical Durability

The chemical durability of a material refers to its ability to resist corrosion and chemical alteration, which usually result from exposure to aqueous solutions. In the case of a waste form, corrosion can result in the release of radionuclides. In a geologic repository, radionuclides may be transported in the groundwater to the biosphere. Chemical durability is usually measured as a corrosion rate, dissolution rate, or leach rate. Leach rate is measured in gm/m2day and commonly varies between 10-4 - ~1 gm/m2day for borosilicate glasses and Synroc ceramics.101

A number of variables affect the corrosion rates of different waste forms. The composition and ionic strength of the leachate, the pH of the leachate, and the temperature are the three most important parameters that affect corrosion rate.102 The flow rate of the leachate, the redox potential, the waste package materials, the waste loading, and radiolysis of the leachate are also factors that affect corrosion rates.103 For ceramics, the production methodology (hot pressing versus cold pressing, for example) can also affect chemical durability.104 For a material such as Synroc, it is important to consider the corrosion behavior of each phase separately. In the case of plutonium immobilization, the leachability of pyrochlore or zirconolite are the most significant, because they are host phases for plutonium.

Workers have done numerous experiments on borosilicate glass corrosion under a wide range of conditions.105 Under experimental conditions of constantly flowing deionized water at 90oC, the range of corrosion rate for borosilicate glass is from 0.5-5 g/m2day.106 Compare this range to long-term corrosion rates of 0.01-0.001 g/M2day.107 The difference in corrosion rate ranges may have to do with silica saturation of the leachate. Under static groundwater conditions, the leachate becomes saturated in SiO2 relatively rapidl