REDUCING STOCKPILES OF EXCESS FISSILE MATERIALS
Bunn, Matthew, and John P. Holdren. "Managing Military Uranium and Plutonium in the United States and the Former Soviet Union." Annual Review of Energy and the Environment 22 (1997): 403-486.
No matter what security and monitoring measures are imposed on the huge stockpiles of excess plutonium and HEU now arising, these materials will continue to pose significant security risks as long as they remain in combinations of forms and locations from which they could be rapidly returned to use in weapons if political circumstances change or security arrangements collapse. Hence, there is a growing international consensus [reflected, for example, in the statement of the 1996 Moscow Nuclear Safety and Security Summit (68)] that, as rapidly as practicable, physical barriers should be erected that would make it substantially more costly, difficult, and time-consuming ever to reuse these materials in weapons— either by transforming the materials physically or by removing them to locations from which their recovery would be physically impractical.
This situation has motivated considerable attention, which began even before the Cold War ended, to the technical options for accomplishing this objective (for early studies see e.g. 73-78). It has become customary to use the term "disposition" for this phase of nuclear weapon-material management, which goes beyond engineered and guarded interim storage and entails physical transformation and/or relocation of the materials with the aim of increasing the difficulty and cost of their recovery and reuse for weaponry.
A great many approaches to disposition have been conceived and, in varying degrees, analyzed in recent years. As described in the 1994 NAS report (8), these include the following classes of options:
- Burning the material as fuel in nuclear reactors. Either currently operating or more advanced types of reactors could be used.31 The fuel cycle could be "once-through" (whereby part of the weapon material is fissioned and part remains embedded in the spent fuel together with highly radioactive fission products), or it could entail fuel reprocessing with recycle of the recovered fissile materials (which, depending on the reactor/fuel combination, the number of cycles, and how the fuel cycle is operated, may be able to fission a higher fraction of these materials than a once-through cycle). The residual (unfissioned) nuclear-weapon materials from these approaches would become part of the radioactive wastes that will be produced in any case from nuclear energy generation and would accompany these wastes to whatever final resting place society chooses for them (which, for the United States, now seems likely to be a mined geologic repository).
- Mixing the material with pre-existing high-level radioactive wastes in an immobilized waste-form for eventual geologic disposal. Large quantities of high-level radioactive wastes have been generated in the course of both nuclear-weapons production and nuclear-energy generation. Surplus fissile materials could be mixed with these— preferably intimately mixed to maximize the difficulty of reseparation, and preferably in a stable matrix that would minimize leakage of radioactive isotopes from the waste package after eventual geologic disposal. The resulting immobilized mixtures of fissile materials and radioactive wastes would then be disposed of in the same final resting places chosen for the rest of the high-level military and civilian wastes.
- Disposing of the material, without mixing it with radioactive wastes, in locations selected to make recovery difficult or impossible. Options that have been mentioned in this category include burying the material in sediments on the deep ocean floor, burying it beneath the Antarctic ice sheet, placing it at the bottom of deep (several kilometers) boreholes in solid rock, diluting it to low concentrations in the open ocean, and launching it into the sun or out of the solar system on rockets.
It would also be possible to combine the third category with the first or second, that is, to dispose of spent fuel or mixtures of fissile materials and high-level radioactive waste in some of the ways mentioned under heading 3, most of which entail a higher level of difficulty of access— and probably higher costs of emplacement— than the mined geologic repositories currently contemplated for disposal of the bulk of military and civilian high-level radioactive wastes. Finally, an approach that resembles category 2 but combines a small amount of new-fission-product creation, mixing with the surplus fissile materials, and geologic emplacement all in one step is to detonate a nuclear bomb in the midst of suitably arranged surplus weapon material underground, resulting in an underground cavity lined with glassified rock formed by the explosion and containing its fission products along with the surplus material.
What criteria should be used to choose among this array of possibilities for fissile-materials disposition? Probably most people would agree that the criteria should include at least the following:
- timing, meaning how quickly the option could be brought into operation and how quickly it could complete processing of the quantities of surplus weapon-usable material that need to be dealt with, which is critical to reducing, as rapidly as practicable, the security risks posed by continued storage of the material;
- the dangers of diversion of the material to weapons use (by the country that initially owned it) or its overt or covert theft for weapons use (by anyone else), which are affected by the magnitudes of the flows and stocks of weapon-usable material and the combination of intrinsic, locational, engineered, and institutional barriers surrounding it, at each step of the disposition process (including the material's final state);
- compatibility with wider arms-control and nonproliferation goals;
- environment, safety, and health (ES&H) considerations at each step in the disposition process, including, for example, occupational and public exposures to radioactivity and radiation in the course of routine disposition operations, safety of disposition operations against accidents (including criticality accidents in material processing and storage, and reactor accidents), and effects on the character and magnitude of the radioactive-waste-management burden faced by society;
- net monetary costs of the disposition operations (after allowing for any saleable products), not only up to the time of emplacement of the material in its final location but also, if applicable, thereafter; and
- the magnitude of the uncertainties attached to all of the foregoing characteristics, including not only technical uncertainties but also uncertainties (especially of timing) arising from institutional complexities and potential difficulties of securing public acceptance.
Much harder than listing the relevant factors, however, is addressing what relative weight should be assigned to them and how to deal with trade-offs among them (and among different aspects of a given characteristic, such as security or environment).32
One of the more important contributions of the NAS study (8) to the disposition discussion, in our view, was in sorting out some of the weighting and trade-off issues by (a) defining standards of sufficiency (answering the question, "How good is good enough?") for some of the factors and (b) specifying which of the other factors should dominate choices among those options that meet these sufficiency standards. More specifically, the NAS report argued that sufficiency in a disposition option entails the following.
- With respect to the final condition of the fissile material, plutonium should end up not significantly easier to recover and reuse in nuclear weapons than is the plutonium in typical spent fuel from civilian nuclear reactors (the "spent-fuel standard").33
- With respect to ES&H aspects, the disposition processes should (a) comply with existing national regulations (and subsequent modifications of these) governing allowable emissions of radioactivity to the environment, allowable risks of accident, and allowable radiation doses to workers and the public from civilian nuclear-energy activities in the country where the disposition operations take place; (b) comply with existing international agreements and standards (and subsequent modifications of these) governing nuclear safety and radioactive materials in the environment; and (c) not add significantly to the ES&H burdens that would be expected to arise, in the absence of weapon-materials disposition activities, from responsible management of the environmental legacy of past nuclear-weapons production and from responsible management of the ES&H aspects of past and future civilian nuclear energy generation.
The NAS report argued further that, among the disposition options meeting the foregoing sufficiency criteria, the choice should be most heavily influenced by considerations of most advantageous timing (soonest start, earliest completion), closest approach to the stored weapons standard (described above) in relation to the quality of materials protection and accounting achievable throughout the disposition process, and smallest uncertainties (technical, institutional, and public-acceptance). Monetary costs, while not unimportant (especially in the cash-short former Soviet Union), should be less important than timing and other security considerations unless the costs differ by margins substantial enough to affect more important criteria, such as the timing and uncertainty of carrying out the undertaking at all.
Since the appearance of the NAS study, the essentials of this approach to the disposition issue, and particularly the spent-fuel standard, have been endorsed by a number of other studies (see e.g. 33) and by the US government (see e.g. 34, 36, 37, 39). Its key ingredients have been endorsed, as well, by Russia and the other major industrialized nations, as reflected in the Moscow summit statement just referred to, which in particular endorsed the spent-fuel standard by calling for surplus weapon materials to be "transformed into spent fuel or other forms equally unusable for nuclear explosives." The statement also called for the application of international safeguards and stringent standards of security and accounting for the nuclear materials, and for heavy emphasis on nonproliferation and environment, safety, and health objectives (68).
Highly Enriched Uranium Disposition
In the case of HEU, it has turned out to be relatively easy to find a disposition method that meets the "sufficiency" conditions for end-point security and ES&H characteristics, could be started with little delay and completed rather quickly, allows for a high standard of materials protection and accounting at all intermediate steps (including its entailing a minimum of transport of weapon-usable material in vulnerable forms), and actually makes rather than costs money. This attractive situation results from the fact that highly enriched uranium can be blended in a technically straightforward way with natural or depleted uranium to produce proliferation-resistant LEU, which is a valuable fuel for reactors of currently operating civilian types.
The 3-5% concentration of U-235 in these LEU fuels is well below the minimum required for making a nuclear explosive and far below the 90% or higher figure preferred by weapon designers. LEU can only be re-enriched to weapon-usable U-235 concentrations, moreover, by prolonged use of technologically demanding and costly uranium-enrichment facilities. Only a dozen or so nations, and no subnational groups, have such facilities; they are not easy for a nation to acquire or conceal, and they would not be easy for a subnational group to take over for long enough to use them for enriching LEU to weapon-usability.
The approach of "blending down" HEU to make LEU is the one that both the United States and Russia have decided to use for their excess HEU stockpiles. The US DOE issued a Record of Decision in July 1996 recording its decision to blend the roughly 175 tons of excess US HEU to LEU (79). In September 1996, DOE published a plan for HEU disposition, under which as much as 85% of the blended LEU would be sold for commercial fuel over 20 years (the remainder being extensively contaminated material that will be disposed of as waste after blending) (80).
The United States has also agreed to purchase LEU blended from 500 tons of Russian excess HEU over a period of 20 years (for the article widely credited with suggesting this idea, see 81). If current prices persist, the value of the deal over that period will be roughly $12 billion. A series of difficult negotiations took place over the period 1992-1996 to reach this agreement, work out the implementation details, settle on prices and delivery schedules, and arrange this material's entry onto the commercial market in a way that would not unduly depress uranium and enrichment prices or provide a basis for successful legal action by producers of uranium and enrichment services (for a critical discussion of US handling of these negotiations, see 82). Most of these issues have been resolved (at least for now), and the deal is moving forward— material has been blended from some 18 tons of HEU already delivered, and a contract was recently signed that resolves disputes over pricing and delivery rates for the next five years. Unfortunately, however, in the spring of 1997 another obstacle arose, involving US legal constraints on the re-export to Russia of natural uranium blendstock provided in payment for the natural uranium component of the delivered material, which led to a Russian decision to halt shipments; as of mid-1997, US officials expected this obstacle to be overcome and deliveries to resume shortly thereafter.
Currently, the HEU deal is for 500 tons of material to be purchased over 20 years. There are strong security arguments for increasing both its size and speed. Russia has indicated informally that it has substantially more than 500 tons of HEU it would be willing to sell. As with the initial deal, additional purchases would help reduce the stockpiles of weapon-usable material in Russia, create an additional incentive for weapons dismantlement, and provide much-needed hard currency— all at zero or modest net cost to the US taxpayer. Arrangements might be reached, moreover, under which the profits from additional purchases might be used to fund high-priority nuclear security objectives, such as upgrading MPC&A or undertaking plutonium disposition (see below). Speeding up the deal would reduce the time during which this material remained in weapon-usable form and bring the other benefits just described more rapidly. Even if the commercial market cannot absorb the material more rapidly, it would make sense to attempt to expand available blending capacity to blend the material as rapidly as practicable. If blending capacity cannot be expanded dramatically, the practicality of other options should be closely examined, including blending the material rapidly to an intermediate level below 20% enrichment, so that it is no longer usable in weapons, or shipping it to the United States as HEU, for blending at a later time (82). The full impact of these options should be considered, however, including the costs of blending before the material can be released on the market and the impact on needed jobs in Russian nuclear cities of shipping the material to the United States as HEU.
Plutonium Disposition: Narrowing the Options
Plutonium raises more difficult issues, and hence our treatment of plutonium disposition in this article will be substantially more detailed. Because virtually all mixtures of plutonium isotopes can readily be used for nuclear weapons, plutonium cannot simply be diluted isotopically, as HEU can, into a form that would require technically demanding isotopic re-enrichment before becoming usable in weapons again.34 Uranium ore and enrichment services are so cheap, moreover, that even "free" plutonium is not economic to use in water-reactor fuel in competition with low-enriched-uranium fuel in a once-through fuel cycle. [An important contributor to this surprising result is the very high cost of fabricating plutonium-based fuel, arising from the special security safeguards and worker-health precautions it requires. See (9).] As a consequence, disposition of surplus weapons plutonium will cost money, not make money, no matter which option is chosen.
The NAS study reviewed the full array of options outlined above for disposition of surplus nuclear-weapon plutonium, applying the indicated sufficiency criteria and examining the options appearing to meet these criteria in terms of the further desiderata of timing; other security characteristics; detailed ES&H properties; costs; and technical, institutional, and public-acceptance uncertainties. The study concluded that, while all plutonium disposition options suffer from drawbacks and significant uncertainties, the two least problematic disposition options are
- fabrication of the plutonium into mixed-oxide fuel (a mixture of plutonium dioxide and uranium dioxide, termed MOX) for use on a once-through basis in a limited number of civilian power reactors of currently operating types (albeit possibly with some modifications to increase the allowable plutonium loading per reactor, thus speeding up the process or reducing the number of reactors needed); or
- vitrification of the plutonium together with high-level radioactive wastes, achieved by mixing the plutonium and fission products from previous military or civilian nuclear energy activities into molten glass to produce glass "logs" of mass, bulk, radioactivity, and resistance to chemical separation of the plutonium comparable to these properties for spent-fuel bundles from civilian reactors.
The NAS study found that both of these approaches would meet the spent-fuel standard; that both could be accomplished within the ES&H sufficiency criteria described above; that they would be comparable to one another in timing, other aspects of security, overall level of uncertainties, and cost; and that both of them would be superior in timing, cost, and uncertainties to all of the other options investigated that could meet the spent-fuel standard and the ES&H sufficiency criteria.
The conclusion that suitable versions of the immobilization-with-wastes option meet the spent-fuel standard deserves some elaboration. That standard refers not to matching any single characteristic of spent fuel but to making it roughly as hard to acquire the material, recover the plutonium, and make a bomb from it as it would be to do the same with the plutonium in commercial spent fuel. This overall difficulty in returning the plutonium to weapons results from the various barriers that characterize typical spent fuel from currently operating commercial reactors— the mass and bulk of the fuel elements, their radiation field, the low concentration of the contained plutonium and the difficulty of separating it chemically from the materials with which it is intermixed, and the deviation of the plutonium's isotopic composition from the ideal for weapons use. Achieving the spent-fuel standard means that the material's characteristics pose difficulties for theft and weapons use of the plutonium that are generally comparable to those associated with typical spent fuel, which itself varies with reactor type and the specific fuel's history inside and outside the reactor; it does not mean that the material needs to be identical, in each category of barrier, to a particular type of spent fuel.
The immobilization option, unlike the MOX/current-reactor option, does not change the isotopic composition of weapon-grade plutonium at all. But because the NAS committee judged that isotopic variations from the weaponeer's ideal are a much smaller barrier to bomb-making than is intimate mixing of the plutonium with fission products (and no barrier at all to actual theft, compared to mass, bulk, and fission-product radioactivity), the committee held that the standard would be met by plutonium- and waste-bearing glass logs whose mass, radioactivity, and difficulty of chemical processing were expected to be generally comparable to those of typical spent fuel.35
In terms of the risk of theft and proliferation, the difference between reactor-grade and weapon-grade plutonium is very modest. For the United States or Russia, weapon-grade plutonium would be somewhat more attractive for reincorporation in their arsenals in the event of a reversal of current arms reductions (since it could be used with high confidence in existing designs without nuclear testing), but the cost and difficulty of recovering the material would be substantial, and both the NAS and subsequent DOE studies concluded that overall, the level of "irreversibility" offered by the reactor and immobilization approaches would be roughly comparable. Perceptions that leaving the material in weapon-grade form offers an option for rapid reversal could be important, however, as discussed below.
Under current US policy, the ultimate fate of the plutonium-bearing waste form from either of these options— spent MOX fuel or plutonium and high-level wastes in glass logs— would be in a geologic repository. Spent fuel and glass logs containing high-level wastes will exist in large quantities and will need to be safely managed and eventually disposed of regardless of what happens to the excess weapon plutonium. The NAS study's conclusion that these two options are the most attractive ones available does not depend on emplacement of the waste forms in a particular repository or by a particular time, or even on emplacement in a repository at all. The key point is that once the weapons plutonium is embedded in spent fuel or waste-bearing glass logs of suitable specifications, it will be approximately as resistant to theft or diversion as the larger quantities of reactor-grade plutonium in commercial spent fuel— and will then represent neither a unique security hazard nor a large addition to the radioactive-waste-management burdens that the spent fuel and immobilized defense wastes would pose in any case.
Nevertheless, both approaches require further study to determine what modifications may be needed to the nuclear-energy and nuclear-waste-management facilities and procedures to ensure that the addition of weapon plutonium would not pose significant problems in relation to accidental chain reactions, worker hazards, and other ES&H issues. For both approaches, it is important to ensure that the characteristics of the plutonium-bearing waste forms and their repositories preclude the residual fissile material's ever reaching criticality in the ground and that addition of weapon plutonium to these wastes does not add significantly to the environmental hazards expected to result from disposal of these wastes (one of the sufficiency criteria).
Subsequent to the publication of these NAS findings, very detailed reexaminations of the full array of plutonium disposition options by the US DOE (34-36, 58, 83) concluded, similarly, that once-through MOX and immobilization with radioactive wastes are the two least problematic plutonium-disposition possibilities and that both would meet the spent-fuel standard. A detailed US-Russian joint study also concluded that these two options are feasible and could meet the spent-fuel standard, though it made no choices of recommended options (37). The reactor-MOX option and the immobilization approach were also identified as the two leading options in the Moscow summit statement and at the subsequent International Experts Meeting on Disposition of Excess Weapons Plutonium. The only significant difference between the more recent official studies and the NAS study in these respects is that "vitrification" has now been replaced with "immobilization" in the description of the mixing-with-wastes option to leave open the possibility that the embedding matrix might not be glass but rather a ceramic or other type of material.
The NAS study's preference for glass— more specifically, for borosilicate glass— is based on this material's having already been relatively well studied in relation to its capacity to contain radioactive wastes and plutonium in a geologic-repository environment and having been the material of choice in most countries for HLW disposal, where there has been substantial industrial experience with its production and handling. This, in the NAS committee's view, means that choosing another glass type or a non-glass embedding material might require substantially more time for testing and licensing before disposition using the immobilization approach could begin. There are, however, some arguments in favor of using a ceramic rather than a glass matrix for immobilizing radioactive wastes, and the question of which is the most appropriate embedding material for such wastes— whether or not they contain plutonium36— has become a more active area of investigation over the past few years (84-89).
All of the other options considered in the NAS study or subsequently either (a) failed to meet one or more of the sufficiency criteria proposed by the NAS study or (b) met or exceeded these criteria, but with longer delays, greater uncertainties, and/or higher costs than the leading candidate options. We now summarize the reasons why specific other disposition options were deemed less attractive, by the NAS study and the subsequent DOE studies mentioned above, than the once-through-MOX and immobilization-with-wastes options (for more detail see 8, 9, 83).
DEEP BOREHOLES . . . Deep boreholes were identified in the 1994 NAS volume as the third option for near-term plutonium disposition most deserving of further study. They also survived DOE's process of screening out unreasonable options, and hence were analyzed in some detail in DOE's studies of the reasonable disposition approaches (34-36). Placing plutonium in very deep boreholes would provide excellent protection against its theft by subnational groups; the protection offered against recovery by the country that emplaced it would be somewhat less, though how it compared to the difficulty of recovering plutonium from spent fuel would depend on the details of the emplacement scheme. Unlike the once-through-MOX and immobilization-with-wastes options, however, the borehole option would achieve its security benefits only once a borehole repository had been approved and licensed and the geologic emplacement accomplished. Given the history of nuclear waste management in the United States, substantial implementation delays, which would prolong the risks of continued storage in directly weapon-usable form for an unpredictable period, seem likely with this approach. Ultimately, it was judged impossible to have sufficient confidence that the borehole option could be accomplished on a reasonable timescale for reliance to be placed on this option, and DOE eliminated it in the final choice of preferred approaches.
ADVANCED REACTORS AND FUELS . . . During the initial flurry of analysis and thinking about plutonium disposition that followed the end of the Cold War, there was considerable enthusiasm in the nuclear-reactor community about the possibility that the plutonium-disposition mission might serve as the rationale for government support for the development and demonstration of new nuclear-reactor types that would combine high burnup capabilities for weapon plutonium with improved efficiency, economics, safety, and/or waste-management characteristics in the nuclear-energy-generation role. In the early 1990s, Congress directed DOE to sponsor a variety of studies of these possibilities by reactor manufacturers (see e.g. 90). The approaches that were studied included advanced light-water reactors using both conventional MOX fuels and advanced fuel types lacking fertile material (91-94), high-temperature gas-cooled reactors (95, 96), and liquid metal reactors (97). Some of them involved the use of plutonium recycle to achieve, ultimately, the fission of a large fraction of the weapon plutonium, and others showed the possibility of achieving burnups in once-through operation that would be considerably higher than those possible in once-through operation in currently operating reactor types with conventional fuels. Studies conducted in DOE national laboratories also explored the use of accelerator-driven subcritical reactors for plutonium disposition and power generation. Versions of this approach using both molten salt and particle-bed cores were examined in some detail, along with fully critical variants lacking the accelerator. (For somewhat critical summaries of such approaches, see 98, 99).
The NAS study, after examining these findings and conducting its own further analyses of many of the advanced-reactor options, concluded that the MOX/current-reactor and immobilization-with-wastes options can achieve the spent-fuel standard more quickly, more cheaply, and with smaller uncertainties than any of the advanced reactor or advanced fuel options and that the advanced options offered no advantages in other aspects of security or in ES&H aspects that would be large enough to justify delaying weapon-plutonium disposition until such options could be developed and deployed.37 An important contributor to this conclusion, in relation to advanced options that could burn up a higher fraction of the plutonium than the 25-40% achieved by once-through MOX use in currently operating reactors, is that there is no great security benefit in pushing beyond the spent-fuel standard for weapon plutonium unless and until one is in a position to similarly reduce the security risks posed by the much larger and still growing quantities of civilian plutonium in spent fuel worldwide. In other words, it makes most sense to move quickly to bring the surplus weapon plutonium to the spent-fuel standard and then ask what further measures society might wish to take to reduce the residual security risks from spent fuel.
SUBSEABED DISPOSAL AND OCEAN DILUTION . . . Burial in the mud layer on the deep-ocean floor, known as subseabed disposal, has long been considered by some to be a leading alternative to mined geologic repositories for the disposal of high-level radioactive wastes (100-102); it could also be considered for the disposal of wastes incorporating excess weapon plutonium. Large areas of the abyssal muds have been geologically stable for millions of years and are thousands of miles from human population centers, and the properties of the mud itself would contain most radionuclides for hundreds of thousands if not millions of years. Emplacement could be accomplished by various methods, including the dropping of appropriately designed canisters from ships on the surface. The canisters in that case would embed themselves tens of meters down in the mud.
Most of the parties to the London Dumping Convention, however, agree that it bans dumping of radioactive wastes not only in the oceans but also in the subocean mud; in late 1996, the parties adopted modifications that made this prohibition explicit. Hence, this approach fails to meet one of the sufficiency criteria in the NAS report: compliance with national and international regulations and agreements. Such an approach would also be likely to generate overwhelming national and international political opposition, creating large uncertainties about whether and when it could be implemented. For these reasons, the NAS report recommends this approach not be pursued unless it is reconsidered for the broader purpose of radioactive waste disposal (8).
Similar arguments apply to proposals for diluting the plutonium in large volumes of the ocean. The basis of this rather counterintuitive idea is a simple calculation showing that mixing, say, 250 tons of weapon plutonium into the combined volume of the world's oceans would yield an average plutonium concentration low enough that current standards would deem it entirely acceptable for drinking water. This approach would considerably exceed the spent-fuel standard, making the plutonium practically irrecoverable. But it would be prohibited by the London Dumping Convention, would evoke substantial political opposition, and would entail the possibility that localized "hot spots" and bioaccumulation in food species could lead to significant human doses (8). The idea of placing the plutonium in the subduction zones at the edges of continental plates, so that it would be carried downward as one plate slips under another, makes even less sense: The motion of the plates is so slow that most of the plutonium would have decayed by the time it moved more than a few meters, and the subduction zones are far less stable than many other parts of the ocean.
LAUNCHING THE PLUTONIUM INTO SPACE . . . At various points in the past, the possibility of launching radioactive wastes into space— into the sun, out of the solar system, or into some other orbit that would have little chance of ever intersecting that of the earth— has also been considered (see e.g. 103). For the weapons plutonium, the costs of such a program would be high, the possibility of failure of one or more of the rockets significant, and the likely public opposition overwhelming; these obstacles would be even greater for the much larger program that would be required to deal with radioactive wastes (8).
VAPORIZING THE PLUTONIUM IN UNDERGROUND NUCLEAR EXPLOSIONS . . . Another proposal is to use underground nuclear blasts for disposition of excess weapons plutonium (104). In one concept, some 5,000 plutonium pits would be arranged around a 50-kiloton nuclear explosive, which, when detonated, would vaporize the pits and 50,000 tons of surrounding rock, all of which would then resolidify underground in a glassified mixture containing rock, plutonium, and a tiny amount of fission products from the detonation. (If safety issues could be resolved, such an approach might even be used with 5,000 assembled nuclear weapons rather than 5,000 pits, avoiding the weapon-dismantlement step, but throwing away the valuable highly enriched uranium also contained in the weapons.) The obvious safety and environmental issues raised by disposing of large quantities of plutonium in an explosively created waste form in a non-engineered repository were never analyzed in detail, as the proposal was easily rejected on other grounds. Since a treaty has now been reached banning underground nuclear explosions, this approach fails to meet the sufficiency criterion of compliance with international regulations and agreements. As the reaction to recent French nuclear testing has shown, moreover, public opposition to such a program would probably be intense, and the plutonium in the rock would remain an extremely rich "plutonium mine" for the country where it was located, failing to meet the spent-fuel standard at least in respect to recovery by the host state (8).
Requirements of the Two Preferred Approaches
To implement plutonium disposition using the preferred approaches will require a number of large-scale facilities, including the following.
- Both options require facilities for converting plutonium weapon-component "pits" to oxides and for processing other types of plutonium to prepare it for disposition.
- The once-through-MOX option requires (a) facilities for fabricating plutonium oxide into MOX fuel and (b) sufficient numbers of reactors capable of safely handling MOX fuel and licensed to do so.
- The immobilization-with-wastes option requires facilities for immobilizing plutonium with high-level wastes (either together in the same immobilized matrix or in separate matrices that would then be intermingled).
Neither the United States nor Russia has large-scale operational facilities for converting pits to oxide, fabricating plutonium fuel, or immobilizing plutonium with radioactive wastes. New facilities will have to be built, or existing facilities modified, for these purposes. In addition, neither country has operating commercial reactors licensed to use MOX fuel. Analysis, testing, and in some cases reactor modifications will be needed to acquire the necessary licenses or license amendments for the use of MOX fuel.
PIT CONVERSION AND PLUTONIUM PROCESSING . . . Both of the preferred approaches require a facility to prepare plutonium for disposition. A prototype facility capable of handling approximately 200 pits per year will be operational at the Los Alamos National Laboratory in early 1998, but a larger facility will eventually have to be built or an existing facility modified for this purpose. New or modified facilities will have to be provided for this purpose in Russia as well. A significant fraction of the US excess plutonium is in forms other than plutonium pits (including various types of impure or contaminated forms, scrap, and the like), which also require processing to prepare them for disposition. Thus, either more than one facility will be required or the processing facility will require the flexibility to handle different types of input materials. This preprocessing step is expected to account for a significant fraction of the total cost of the disposition mission, and it is an important factor in determining when plutonium disposition on a substantial scale could begin (34).
THE ONCE-THROUGH-MOX APPROACH . . . Nearly two dozen light-water reactors (LWRs) around the world are already using MOX fuel made from plutonium reprocessed from civilian spent fuel (105). While some modest differences are associated with using weapons plutonium in such fuel (resulting from both isotopic and chemical factors), the MOX option can be considered essentially a demonstrated technology. Typically, the mix of uranium and plutonium oxides in MOX fuel for LWRs contains 3-5% plutonium.38 Most of the LWRs now using MOX use this fuel in only one third of their reactor cores (with the remaining two thirds containing traditional LEU fuel); this minimizes the change in core neutronic characteristics (and the resulting increase in requirements on reactor control systems) resulting from the different properties of plutonium and uranium (9, 105).
Using MOX in a larger fraction of the reactor core would allow a faster disposition campaign (if sufficient MOX fuel fabrication capability were available) or the use of fewer reactors. Three operating reactors in the United States (the System-80 reactors at Palo Verde) were specifically designed to be capable of handling MOX in 100% of their reactor cores, and recent studies by reactor vendors and DOE have concluded that many other US reactors would be capable of handling MOX in 100% of their reactor cores with only modest modifications while remaining within their existing licensed safety margins (34, 106-109). License amendments would be required for use of either one-third or full-core MOX, and the complications would be greater in the case of full-core MOX. An approach under active consideration, therefore, is to begin with one-third MOX cores and move to full-core MOX later, as the relevant issues are resolved (34).
A typical 1000-MWe LWR using MOX fuel with 4% plutonium in one third of its core, operating at a capacity factor of 75%and irradiating the fuel to 42,000 megawatt-days per metric ton of heavy metal (MWd/MTHM) would absorb 275 kg of weapon plutonium per year;39 thus, under these circumstances, some 180 reactor-years of operation would be required to absorb and irradiate a nominal 50 tons of weapons plutonium. The same reactor using MOX in 100% of its core would absorb about 830 kg of weapon plutonium per year. Consuming 50 tons of excess plutonium in 20 years of operation would require 9 reactors of this type using one-third-MOX cores or three such reactors using full-MOX cores. In either case, supporting the mission with these time parameters would require a pit-conversion facility capable of handling 2.5 tons of plutonium per year and a MOX plant capable of fabricating about 63 MTHM of MOX each year.
Canadian deuterium-uranium (CANDU) reactors could also be used for plutonium disposition; this approach was analyzed in the NAS report and subsequent DOE-sponsored studies and has since been endorsed by the Canadian government (34, 110, 111). CANDU reactors are believed to be capable of handling 100% MOX cores without significant changes to their control systems. Using MOX based on the existing fuel design, with an average loading of 2.2% plutonium, a single 825-MWe CANDU reactor (such as those at the Bruce A station, being considered for this mission) could irradiate over 1.4 tons of plutonium per year to a burnup of 9700 MWd/MTHM. Using the more advanced fuel design now undergoing testing would allow higher average plutonium concentrations of 3.4%; if the burnup were left the same, this would increase the throughput to over 2.2 tons per year, but if the burnup were increased to take advantage of the energy potential of the fissile material in the fuel, throughput would actually decline slightly, to 1.25 tons of plutonium per year (34). Thus, between 23 and 40 reactor-years of operation would be required to irradiate 50 tons of weapons plutonium. Accomplishing this in 25 years would require a pit conversion plant capable of processing 2 tons of plutonium per year and a MOX plant capable of producing 60-90 MTHM of MOX per year (depending on the plutonium loading in the MOX).
In the United States, several dozen reactors, far more than required, have sufficient licensed reactor lifetimes remaining to participate in the plutonium disposition mission. Although some US reactors may shut down as utility deregulation takes hold if their utility owners do not succeed in operating them in a way that is cost-competitive with other sources of power, the number of reactors likely to remain operational will still be far more than sufficient. The number of CANDU reactors in Canada is also more than sufficient for the mission. In Russia, by contrast, only seven modern VVER-1000 LWRs are operational, and there is some doubt as to whether the older units of this type could practically be converted for MOX use; if all seven could be used, but they were limited to one-third cores, it would take more than their remaining licensed lifetimes to carry out disposition of 50 tons of excess plutonium. An additional 11 VVER-1000 reactors, supplied with fuel by Russia, are operational in Ukraine, however; Russia apparently already intends to supply these reactors with MOX fuel made from civilian plutonium originating from reprocessing of Ukrainian fuel, if Russia eventually builds a MOX fuel fabrication plant that could be used for this purpose (37). With one-third cores, these 18 reactors combined could carry out disposition of even as much as 100 tons of excess plutonium in less than 25 years. If the Russian VVER-1000s could be modified to handle full MOX cores at reasonable cost, five of these plants could carry out disposition of 100 tons of excess weapons plutonium in 25 years, without requiring the Ukrainian plants.40
Neither the United States nor Russia has an operational industrial-scale MOX fuel fabrication plant, though such plants are operational in several other countries. The time required to provide such a capability is a key limiting factor on when the once-through-MOX option could begin; DOE estimates that a domestic facility will not begin producing MOX assemblies until 2007. An earlier start could be made on disposition in both the United States and Russia by contracting with existing European facilities for fabrication of initial lead test assemblies and possibly even the first reactor cores, as recommended in the Bilateral Commission reports. Careful attention would have to be paid to the safety and security issues involved in international plutonium transport, however, and the associated political difficulties (11, 12, 34).
The NAS study estimated that disposition of 50 tons of weapons plutonium in existing US LWRs, using modified existing facilities for pit conversion and fuel fabrication, would have a net discounted present cost (at a 7% discount rate) of about $0.5 billion in 1992 dollars, not counting the cost of fees utilities are likely to charge the government for "irradiation services." The equivalent cost for the CANDU approach was estimated at just under $1.0 billion (primarily because the unenriched CANDU fuel the MOX would replace is very cheap, making the gap between the usual cost and the cost for MOX fuel very large). 41More recent DOE studies, using 1996 dollars and 5% discount rate, estimated the costs for similar approaches at just over $1 billion for existing LWRs and nearly $1.7 billion for the CANDU alternative. Costs in Russia are much more difficult to estimate: Capital costs of the required facilities are likely to be in the range of several hundred million dollars. DOE currently estimates that either the LWR or CANDU approaches could begin in 2007-2008, and irradiation of 50 tons of excess weapons plutonium could be completed in 2021-2022 (34).
Since the MOX option is largely technically demonstrated, the principal uncertainties facing its implementation in the United States are political and institutional; it may ultimately prove to be very difficult to acquire the necessary political approvals and licenses to implement the MOX option in the United States. The controversy that has already begun over the possible use of plutonium in US commercial reactors is described below.
THE IMMOBILIZATION-WITH-WASTES OPTION . . . Both the United States and Russia are currently immobilizing high-level wastes in glass, but neither of these operations is designed to handle substantial quantities of plutonium safely. (Compared to immobilizing high-level wastes without plutonium, plutonium immobilization requires dramatically different security and safeguards approaches, new measures to prevent accidental nuclear chain reactions as the waste form is being produced, and careful design of the waste form to ensure appropriate performance after eventual disposal in a geologic repository.) Two approaches are being considered to provide the necessary capability: (a) building new immobilization facilities designed to handle both plutonium and fission products or (b) immobilizing the plutonium and fission products separately using existing facilities and intermingling the immobilized products to create similar barriers to theft and recovery of the material.
A particular version of the second approach, known as "can in canister," is now judged to be the leading contender in DOE's immobilization program. In one recent variant of this concept, the plutonium would be immobilized without fission products, in small glass melters or ceramic production facilities installed in existing plutonium-handling glove-box facilities (such as those at the Savannah River Site, where DOE's principal HLW immobilization operation is underway). The resulting "pebbles" of immobilized plutonium would be placed into aluminum cans, which would be arrayed in the large glass canisters used to contain immobilized HLW. These canisters would then be filled with molten glass containing HLW (as they normally would be); the aluminum cans would melt, allowing the molten glass to flow in amongst the plutonium pebbles, intimately mixing the plutonium and the radiation barrier. (Concerns about the impact of the aluminum from the cans on the chemical properties of the surrounding glass are motivating consideration of other approaches to achieve similar results.) In the end, the plutonium would be contained within a 3-m, 2-ton glass log that would generate a radiation dose rate of 400-1000 rads per hour at one meter (decreasing to 200-500 rads/hour after 30 years), comparable to the dose rate from a spent fuel assembly. (For a comparison of current estimates of the physical characteristics of different disposition waste forms, see 36). The advantage of this approach is that existing facilities can be used, reducing costs and delays, and the glass or ceramic to contain the HLW, or to contain the plutonium, can each be optimized for its particular purpose.
DOE estimates that immobilization using such a can-in-canister approach would have a net discounted present cost of just under $1 billion (1996 dollars, 5% discount rate), roughly the same as the costs of using existing LWRs (though utility fees will eventually increase the latter costs somewhat). Immobilization using new facilities designed to safely mix plutonium and HLW in a single immobilized product, with the new facilities built as adjuncts to existing operations, is estimated to have a net discounted present cost of over $1.8 billion, because of the high cost of building new facilities designed to handle intensely radioactive HLW. (If the capability to handle plutonium were built into new HLW handling facilities that have to be built in any case, such as the planned HLW melters to be built at Hanford, the net additional cost would presumably be reduced significantly.) DOE estimates that immobilization using the can-in-canister approach in modified existing facilities could begin in 2006 (or a few years earlier if already available plutonium oxides are used to begin), and immobilization of 50 tons of excess weapons plutonium could be completed after 9-10 years of operation, while immobilization using the new adjunct melter approach could begin in roughly 2009 and be completed after a similar period of operation (34).42
Like the United States, Russia is already vitrifying high-level wastes. While Russia could probably undertake plutonium immobilization on a schedule similar to that of the US if it decided to do so (and Russian experts are carrying out technical analyses of this approach with US funding), Russia is unlikely to dispose of any large fraction of its excess weapons plutonium, which Russian officials generally view as a national asset to be used for energy production.
Because plutonium has never been immobilized for disposal on a large scale before, this option presents significantly greater technical uncertainties than the demonstrated MOX option. Political and institutional difficulties, however, are likely to be somewhat less for the immobilization approach. The NAS report judged that the overall uncertainties facing the two approaches were roughly comparable.
The second volume of the NAS study, after reviewing the issues and uncertainties in the two preferred approaches in detail, recommends that implementation of both options be pursued in parallel, as quickly as possible, "because it is crucial that at least one of these options succeed, because time is of the essence, and because the costs of pursuing both in parallel are modest in relation to the security stakes" (9, p. 14). The Bilateral Commission reports make the same recommendation (11, 12).
After several years of intensive studies and analysis, in early December, 1996, the US government announced that it had chosen this "dual track" approach as its preferred alternative for disposition of US excess weapons plutonium. This decision has provoked considerable controversy, described below. Nevertheless, on January 14, 1997, with the personal approval of the President, DOE confirmed the dual-track approach in its Record of Decision on plutonium disposition (58). DOE expects the net discounted life-cycle cost of implementing the dual-track approach to be roughly $1.3 billion, only modestly more than implementing either the reactor or the immobilization tracks by themselves (34). DOE is now moving to carry out the necessary preparations for implementing both tracks as rapidly as practicable; the Record of Decision specifies, however, that the extent to which "either or both" of these technologies will actually be implemented will be decided in the future (58).
Russian Plutonium Disposition and International Cooperation
US government deliberations concerning plutonium disposition have devoted considerable attention to trying to ensure that disposition of Russian excess plutonium goes forward in parallel. Disposition of US excess plutonium is closely linked to disposition of Russian excess plutonium. It is highly unlikely that the US Congress will agree to finance the significant costs of disposition of US excess plutonium unless Russia has committed to carry out disposition of its excess plutonium; and Russia has already formally indicated that it will not carry out disposition of its plutonium unless the United States is doing so as well.
The situations in the United States and Russia, however, are very different. While the United States has decided not to pursue a civilian fuel cycle based on plutonium reprocessing and recycling and sees its excess weapons plutonium primarily as a security hazard, Russia continues to plan to implement a plutonium-recycling fuel cycle and sees both weapons plutonium and civilian plutonium as essential parts of that plan. Indeed, alone among the major nuclear powers, Russia still hopes to build commercial-scale fast-neutron breeder reactors in the near term (using a design designated the BN-800, similar to the currently operating BN-600 reactor), and Russian nuclear officials have advocated using the excess weapons plutonium as fuel for these reactors once they are built— though funding to complete these reactors is unlikely to become available for many years.
Perhaps the most fundamental problem with disposition of plutonium in Russia is money. As noted above, capital investments of hundreds of millions of dollars would be needed to provide industrial-scale pit conversion, MOX fabrication, and/or vitrification facilities in Russia— a figure that would grow into the billions if new reactors such as the BN-800s had to be built as well. Given Russia's current economic circumstances, it appears very likely that if these investments are to be made in the relatively near term, the international community will have to help finance them. For better or for worse, it appears unlikely that the United States will agree to pay all the costs of its own disposition and Russia's, so other countries may have to participate as well.
One obvious approach to this issue would be for the United States to simply purchase the Russian plutonium, as it is doing in the case of HEU. The material could be purchased and either brought to the United States or some other country for disposition, or disposition could be carried out in Russia. Russian officials have sometimes argued that since plutonium has the same energy value as HEU, it should have the same market value (a notion that is contradicted by the far higher cost of extracting the energy from the plutonium, as described above); by that standard, using the estimated prices that pertain in the HEU purchase agreement, the purchase of, for example, 100 tons of plutonium would cost approximately $2.4 billion. Presumably after purchasing the material, the United States would then be responsible for its disposition, at a substantial additional cost. While we regard these figures as small by comparison to the security stakes, the difficulty of gaining appropriations for such sums in the current environment of budget stringency has been considered so daunting that the US government has never seriously contemplated this approach. Moreover, the political difficulties of importing 100 tons of Russian plutonium into the United States would be severe, and the argument for spending a large sum to buy the material only to leave it in Russia for disposition would be difficult to make. Ashton Carter and others have proposed an approach that is similar to a purchase in some respects, in which an international fund would be established that would pay Russia (and the United States) to place their plutonium in internationally guarded and monitored storage facilities. While this approach would have the novel and useful feature of establishing international, rather than solely national, guarding of the facilities where this particular plutonium was stored, it has never been seriously considered within the US government— perhaps because Russia has already agreed to place its excess plutonium in a storage facility being built with US help and has agreed in principle to international monitoring (though not guarding) of this facility without being paid to do so (see 8, 82).
Thus, the focus of discussion of disposition of Russian excess plutonium has primarily been on options in which the plutonium would remain Russia's and disposition would occur either completely within Russia or at most making use of reactors in other countries after the plutonium was already fabricated into fuel (as in the Canadian and Ukrainian cases mentioned above). The United States, Russia, and other countries have been working slowly to attempt to find a mutually agreeable approach. In their January 1994 summit statement, President Yeltsin and President Clinton directed their experts to conduct a joint study of the options for disposition of excess weapons plutonium. This government-level study, completed and published in September 1996, covers a range of different options, providing assessments of technical feasibility, cost, schedule, nonproliferation impact, and other matters. While the study did not make specific recommendations, the forum it provided for ongoing discussion of plutonium disposition issues has proved to be an important channel of communication (4, 37). Following up on the joint study, US and Russian experts are now conducting joint analyses and tests of key technologies related to MOX, immobilization, and conversion of plutonium "pits" to oxide. To date, however, this channel has been limited almost exclusively to discussions of technical issues related to plutonium dispositi