This chapter describes each of the reactor options examined by the panel. For each option, we offer:
a description of the technology and its development status, including how the technology would be used to process plutonium;
a description of the factors affecting timing of the option, including:
reactor capacity and throughputs,
fuel fabrication, and
licensing and public acceptance issues;
a description of issues related to safeguards and security during the processes required for the option;
a description of issues related to accessibility for use in weapons of whatever plutonium would remain in the spent fuel;
a discussion of economic issues;
a discussion of environment, safety, and health (ES&H) issues; and
a discussion of other issues that may be important to policy-makers in choosing between options.
We begin by discussing several options involving the use of current-generation reactors, then turn to more advanced reactor systems. The description of the first option—the use of light-water reactors of existing designs—will be the most detailed, both because it is an option of particular interest and because
other reactor options can be considered in part by comparison to that basecase. More detailed comparative assessments of the options based on criteria related to security, cost, and ES&H can be found in Chapter 6.
U.S. PLUTONIUM IN CURRENT-GENERATION U.S. LIGHT-WATER REACTORS
Description of Technology and Status
Light-water reactors (LWRs) are the most mature of any of the proposed burners of weapons plutonium (WPu). Over 100 LWRs are operating in the United States and about 400 worldwide. LWRs have over 4,000 reactor-years of operation. They supply almost 75 percent of the electricity consumed in France and about 22 percent of the electricity generated in the United States.
U.S. LWRs use low-enriched uranium (LEU) fuel. Plutonium in the discharged fuel is not reprocessed for recycle. In Sweden and the United States, geologic repositories are being designed primarily for long-term disposal of spent fuel discharged from uranium-fueled LWRs. Other countries are considering such a direct-disposal fuel cycle as well.
It is also possible to fuel LWRs with mixed-oxide (MOX) fuel, which combines plutonium dioxide and natural or depleted uranium dioxide as a PuO2-UO2 mixture. Work on such MOX fuels for LWRs has a long history. The U.S. Plutonium Utilization Program began in 1956, and was soon followed by related work in several European nations and Japan. The development effort was motivated in part by the potential for fuel-cycle economies perceived at the time and also as a means to reduce the consumption of uranium ore. It focused on the reprocessing of LWR discharge fuel to recover and recycle the plutonium and uranium. Several tests of partial core loadings of MOX fuel were conducted in U.S. LWRs during the 1960s and 1970s. Although plutonium recovered from LWR fuel was used in these tests, the results are generally applicable to MOX made from WPu.1
In 1963 Belgium used a partial loading of MOX fuel in its BR-3 pressurized-water reactor (PWR). After many years of experimentation, by 1986 Belgium irradiated a core with a 70-percent MOX loading. Belgium provides a significant fraction of the world's currently operating MOX fabrication capability and is considering an expansion of its MOX fabrication plant. Its MOX fabrication services are marketed in conjunction with France by the MELOX consortium. Two Belgian LWRs are licensed to burn MOX fuel.
Germany tested and demonstrated MOX fuel in LWRs from 1968-1977 and began commercial use of MOX fuel in LWRs in 1981. Seven reactors in
See Chapter 2 for a discussion of the differences between WPu and reactor plutonium (RPu). Details on the U.S. Plutonium Utilization Program are found in USNRC (1976).
Germany are using MOX, five others have a license to do so, and six more have submitted license applications (Thomas 1994, Wilcox 1994). A modest-scale MOX fuel fabrication plant at Hanau operated for several years before losing its license, and a larger plant was nearly completed before encountering licensing difficulties that have so far (late 1994) prevented it from operating.
France began fueling PWRs with MOX fuel on a commercial basis in 1985, and is building a substantial MOX fabrication facility. Sixteen French reactors are licensed to use MOX, and seven of them were doing so as of late 1994 (Nigon and Golinelli 1994). Japan is considering use of MOX fuel in roughly 10 LWRs by approximately the year 2000 (Yamano 1994) and is constructing a substantial MOX fabrication facility and a large commercial reprocessing plant for recycle of plutonium as MOX fuel. Britain, while not having a domestic MOX use program, is building a large MOX plant for foreign customers to complement the reprocessing services it already offers.2
Government-funded work on commercial fuel reprocessing and plutonium recycle in the United States was terminated by presidential directive in the mid-1970s, reflecting concerns that worldwide commercial plutonium recycle might stimulate the proliferation of nuclear weapons, and was also uneconomical (see Carter 1977). While the Reagan and Bush administrations did not take a similarly negative view, U.S. industry has concluded that the high cost of a commercially owned U.S. reprocessing plant makes near-term deployment of reprocessing and plutonium recycle uneconomical in the United States. For that reason, there are now no commercial facilities in the U.S. for reprocessing or for MOX fuel fabrication, and there are no commercial reactors in the United States licensed to use MOX fuel.
From the considerable world experience there is a mature technology adequate to implement the use of WPu as MOX fuel in LWRs. The simplest concept is once-through irradiation of MOX fuel to burnups similar to those used with uranium fuel in LWRs. No reprocessing would be required, and the rate-limiting step would probably be MOX fuel fabrication. Even a small subset of U.S. commercial LWRs would suffice to absorb the nominal 50 tons of excess WPu in this way at the WPu-MOX fuel fabrication rates likely to be attainable.
The following sections present specifics on the technical issues involved in using MOX fuel in LWRs and the rates at which various LWR configurations could process the WPu.
Technical Issues in LWR MOX Fueling
Early plans for MOX fueling in LWRs, in the mid-1960s to mid-1970s, were to reprocess all fuel discharged from a given reactor and to recycle all of
As of late 1994, the target opening date for this plant was 1997 (Wilcox 1994). For more information on civilian plutonium fuel programs, see NAS (1994, Appendix B) and Albright et al. 1993.
the recovered plutonium as MOX fuel for subsequent reloads for that reactor. This was known as self-generated recycle. The neutronic characteristics of LWRs are such that the plutonium produced would be sufficient to fuel about a third of the reactor core when steady-state recycle was reached. The remainder of the core would continue to be fueled with LEU.
The one-third core loading was also thought at that time to be a practical limit for most of the LWRs that had been designed on the basis of uranium fueling. There are several reasons why the fraction of a reactor core that can utilize plutonium fuel without compromising safety margins may be limited, unless the reactor has been designed or modified specifically for plutonium fuel use: the lower delayed-neutron fraction in plutonium fuel as compared with uranium fuel, the higher average neutron energy in plutonium fuel, and the 0.3-electron volt neutron-absorption resonance of plutonium-239 combine to put extra demands on a reactor's reactivity-control systems; the more energetic neutron spectrum and higher gamma-ray flux in plutonium fuels increase radiation damage and thermal stresses in reactor internals; and the higher radioactive-decay energy of plutonium fuels puts greater demands on post-shutdown (including emergency) core-cooling capabilities (see Chapter 2, “Some Differences Between Plutonium- and Uranium-Based Fuels"). For these reasons, and because self-generated recycle could only provide enough plutonium for one-third of the cores of the reactors involved, the designs of most present LWR cores were not intended to offer the capability to burn plutonium fuel in more than one-third of the core.
For disposition of excess separated plutonium, however, using MOX in all of the reactor core rather than only one-third would be highly desirable, as it would reduce the number of reactors or the time required by a factor of three. This could substantially decrease the amount of transportation and the number of sites required to process the excess plutonium, reducing risks of theft during transport and reducing the political and licensing liabilities of involving more of the reactor industry in the plutonium-processing operation. There are also some technical incentives to move in the direction of full-MOX cores. Because the MOX fuel rods have a higher fission cross-section than do the all-uranium rods, the two different kinds of fuel rods must be carefully distributed within the reactor core to avoid local overheating. Also, the reactivity of MOX fuel tends to change more rapidly during an irradiation cycle than that of uranium fuel. There is greater spatial self-shielding of neutron flux in fuel assemblies composed of MOX fuel rods. Design problems are aggravated by the presence of both uranium and MOX fuel rods in the same core. Thus, a reactor core fueled entirely with MOX fuel would be easier to design and to program for fuel reloads.
During the 1970s the U.S. nuclear industry did envision an alternative to self-generated recycle of plutonium with one-third MOX cores. Some utilities planned to dedicate some of their LWRs to operate as "plutonium burners," to be fueled entirely with MOX fuel. An all-MOX plutonium burner would receive
make-up plutonium produced in other uranium-fueled LWRs. The all-MOX plutonium burner would be designed with the extra control absorbers needed for reactivity control, and other modifications would be needed to compensate for the increased fast-neutron and gamma fluxes, the increased decay heat, and the smaller delayed-neutron fraction.
Such a plutonium-burner PWR was designed by Combustion Engineering (Shapiro et al. 1977, Anderson and Klinetob 1981). Called the System-80, this reactor was designed for flexibility to use large plutonium loadings, up to and including a full loading of MOX fuel. As compared to a PWR designed only for uranium fueling, the System-80 plutonium burner has additional control rods and drives and increased boron concentration in the reactor coolant. The cooling systems for the reactor and spent fuel storage are sized for higher long-term decay heat. Core internals are designed to improve cooling and reduce thermal stress that would otherwise result from increased gamma-ray heating. The thickness of the core support barrel is increased to mitigate the increased fast-neutron flux.
Three System-80 reactors, each generating 1,300 megawatt electric (MWe) (3,750 megawatts-thermal; MWt), are in operation in the United States at the Palo Verde Nuclear Generating Station in Arizona. Four 1,000-MWe units are being constructed in Korea, including a forerunner of the evolutionary System-80+ design. One of the two uncompleted reactors of the Washington Public Power Supply System (WPPSS), designated WNP-3 (Washington Nuclear Project), is a System-80 reactor. This reactor is 75 percent complete and has been kept in mothball status.
In addition, recent studies by the major U.S. reactor vendors funded by the U.S. Department of Energy (DOE) have concluded that contrary to past expectation, many (though not all) existing LWRs (both PWRs and boiling-water reactors) could use WPu MOX in 100 percent of their reactor cores with little if any modification.3 Existing margin in the control capabilities of these reactors would, it is reported, allow them to operate within existing safety envelopes with 100-percent MOX cores.
In a number of the existing-reactor cases, however, the maximum safe enrichment of plutonium in the fuel would be lower than it would be in new reactors designed specifically for plutonium use, potentially increasing total fuel fabrication costs and the time required to carry out the plutonium disposition mission. It appears likely that existing reactors could handle higher plutonium loadings if modifications were made conceptually similar to those previously believed necessary to achieve full-core operation at all. This issue—including
the trade-offs between the costs and time required to modify reactors and the extra costs and longer times associated with using fuel of only modest enrichment—has not yet been studied in detail but should be.
The importance of these recent studies is considerable, as they suggest that the job of plutonium disposition in existing U.S. LWRs—one of the main options recommended in this report and the report of the parent committee—would be significantly easier than was believed at the time the parent committee's 1994 report was written. It should be noted, however, that these analyses became available late in the panel's deliberations, and the panel has not been able to review them in detail. The panel urges that they be reviewed by the U.S. Nuclear Regulatory Commission (NRC) to ensure that the conclusion that existing LWRs could use full cores of WPu MOX without compromising existing safety margins is correct.
Whatever the conclusion as to whether significant modifications would be necessary, it appears clear that existing U.S. LWRs could be adapted for full-MOX cores. Because of the considerable worldwide experience with MOX fueling in LWRs, the panel judges the technical uncertainty of the LWR MOX option—using either one-third or full-MOX cores—to be low (by comparison to other reactor options for use of plutonium fuels).
Reactor Throughput, Once-Through Fuel Cycle
The rate at which plutonium could be processed in a once-through MOX fuel cycle is given by the product of thermal power, capacity factor, fraction of the core that is MOX fuel, and plutonium weight fraction in fresh MOX fuel, divided by the average fuel exposure at discharge. Consider, for example, the reactor characteristics once contemplated for self-generated plutonium recycle (Shapiro et al. 1977, Hebel et al. 1978): assuming a 1,000-MWe PWR at 34.2-percent thermal efficiency, a capacity factor of 0.70, a one-third core of MOX fuel, a fuel exposure of 30,400 megawatt-days per metric ton of heavy metal (MWd/MTHM), and an initial concentration of 2.5 weight percent plutonium (Anderson and Klinetob 1981), the yearly amount of plutonium supplied as make-up fuel would be 205 kilograms (kg). Assuming a once-through fuel cycle, 50 tons of WPu could be processed by 8.1 such reactors given a nominal operating lifetime of 30 years.
As noted, using MOX in 100 percent of the reactor core, rather than only one-third, would reduce by threefold the number of reactor-years required to irradiate a given initial quantity of excess WPu. Full-MOX reactors comparable to the System-80 (that is, with a capacity somewhat larger than the 1,000 MWe assumed above), operated at 75-percent capacity factor with a 100-percent MOX core, with an initial plutonium content of 2.5 percent by weight and average burnup of 31,000 MWd/MTHM, would process 828 kg of WPu per reac-
tor-year, thus processing 50 tons in 60 reactor-years (for example, two reactors operating for 30 years) (Hebel et al. 1978).
The initial concentration of plutonium in the MOX reloads can be increased beyond 2.5 weight percent, particularly by adding more burble poisons. ABB-Combustion Engineering (ABB-CE), for example, has suggested that in their System-80+ system, a loading of 6.8 weight percent WPu, compensated by the addition of erbia to the fuel (which also has the advantage of smoothing reactivity over the fuel's life in the reactor), would be attainable. In the existing System-80 reactors, however, which have somewhat fewer control rods, ABB-CE estimates a maximum loading of 4.5 percent (ABB-CE 1994).
Estimates of the average enrichment of WPu that could be used safely in full-MOX cores in existing LWRs other than the System-80s vary. Westinghouse concludes that many existing PWRs could use full-MOX cores with an average enrichment of 4.5 percent, comparable to the enrichment possible in the System-80s. ABB-CE and General Electric conclude that their non-System-80 reactors could handle full-MOX cores with enrichments in the range of 3 percent (ABB-CE 1994, GE 1994, Westinghouse 1994). While no substantial reactor modifications are estimated to be required, the various vendors do suggest measures such as including burble absorbers in the fuel, dissolving increased quantities of boron in the cooling water, or changing the material used in the reactor control rods.
What do these figures mean for how fast the plutonium disposition mission could be accomplished? A 1,200-MWe PWR using a full-MOX core with a plutonium loading of 4.0 percent, and a burnup of 42,000 MWd/MTHM would use just under a ton of WPu each year, requiring roughly 50 reactor-years for disposition of the nominal 50 tons of excess material. If a System-80 reactor with modifications could operate with an initial plutonium content of 6.8 percent, and the same average burnup, this reactor would process about 1,700 kg of WPu per reactor-year. so that one such reactor operating for 30 years would suffice to process 50 tons (see Table 6-1).
As can be seen, the use of high plutonium loadings significantly increases the rate at which the WPu can be processed in a given reactor, even when the burnup is also increased. As described in detail in Chapter 6, moreover, use of such high enrichments reduces the net cost of the operation by reducing the number of kilograms of MOX fuel that need to be produced and increasing the energy value of each kilogram of that fuel.
The use of such high plutonium loadings, however, does create some technical issues, similar in some respects to those involved in moving from one-third to full-MOX cores. In general, burble absorbers must be added to hold down the reactivity. These absorbers (erbium oxide, in the case of the ABB-CE proposal) also help preserve negative temperature coefficients of reactivity for the moderator and fuel. Because of the resonance in the neutron absorption cross-section of erbium-167 at energies just above the thermal resonance of plutonium-239
(Pu-239), erbium can counteract the tendency of Pu-239 to contribute to positive reactivity coefficients. Other materials such as gadolinium, dysprosium, or boron may be equally effective in this role, depending on the specific application. Because the effectiveness of the burble absorbers decreases with increasing burnup, as the reactivity of the fuel declines, they contribute to smoothing out reactivity changes during the fuel cycle.
Higher plutonium-loading MOX fuel will also have a higher fissile plutonium content when removed from the reactor than would ordinary LEU spent fuel. This fact will require careful attention to long-term criticality issues in preparing such fuel for geologic disposal. (See discussion below and in Chapter 6.)
To summarize, the existing System-80 reactors were designed from the outset to handle full cores of plutonium fuel. Analyses by the vendors indicate that many other existing reactors could also use full cores of WPu fuel without substantial modifications to the reactors. Further analysis and review would be required to assess this conclusion. NRC review of any proposal to use plutonium fuels in U.S. reactors is likely to be intensive, and NRC review could lead to a requirement for reactor modifications. But even if significant modifications do turn out to be necessary, the panel believes that a variety of U.S. LWRs could be adapted to handle full-MOX cores safely, with sufficient enrichments to carry out the mission in a small number of reactors.
Providing adequate plutonium processing and MOX fuel fabrication capability would be an important pacing factor for processing excess WPu in U.S. LWRs.
Plutonium pits would have to be shipped from Pantex (located near Amarillo, Texas), where no plutonium processing capability yet exists, to a site capable of disassembling the pits and converting the resulting metal to plutonium oxide. No facilities for carrying out the pit-processing operation on the required scale are currently operating, but facilities at Savannah River (South Carolina), Los Alamos (New Mexico), Hanford (Washington), and possibly elsewhere could be modified for this purpose—and new technologies for efficient pit conversion are being developed at the national laboratories. Using a planning figure of 4 kg of plutonium per pit (NAS 1994), processing 50 tons of plutonium in 30 years would require a capability to process more than 400 pits per year. Processing the plutonium more quickly would require correspondingly larger capabilities.
There are no operating MOX fabrication facilities in the United States. LEU fuel fabrication facilities cannot readily be modified for this purpose because of the much higher radiotoxicity and safeguards requirements of plutonium fuel. Processing 50 tons of plutonium in 30 years, at a loading of 2.5 percent by weight plutonium in the fuel, would require a fabrication facility with a 67-
MTHM/yr capacity; at a loading of 6.8 weight percent, the required capacity would be 25 MTHM/yr.
There is an existing, nearly completed MOX fabrication facility at the federal government nuclear site in Hanford, Washington, known as the Fuel Materials Examination Facility (FMEF). This facility was built in the late 1970s and early 1980s to produce fuel for the liquid-metal Fast Flux Test Facility (FFTF). It was never operated. The panel has received estimates (which may be optimistic) that this facility could be modified to produce 50 tons of LWR fuel per year or more (containing roughly 3 tons of WPu, at a 6- to 7-percent enrichment) while meeting current safeguards and ES&H standards, within roughly five years of receiving a go-ahead, for a cost in the range of $75-$150 million.4 Funds would be needed to replace or upgrade existing, outdated process computers; upgrade older facility systems such as fire protection and waste handling; upgrade other facilities such as security and radiological protection to meet current federal, state, and DOE requirements; prepare safety analysis and safety and compliance documentation; modify the facility for LWR rather than liquid-metal reactor (LMR) fuel; and provide higher throughput than was originally planned.
Alternatively, a new plutonium fuel fabrication facility could be built. Estimates provided to the panel (which appear to be optimistic) indicate that such facilities could be built for between $400 million and $1.2 billion, depending on their capacity. Siting, designing, building, and licensing such a facility would probably require a decade or more. These cost and schedule issues are discussed in more detail in Chapter 6.
Reactor and Institutional Options
Many specific variants of MOX use in LWRs in the United States can be imagined. The reactors used could be existing, partly completed, or newly built for this purpose. Fuel fabrication could rely on partly completed or modifiable facilities or new ones. The relevant facilities could be government-owned and government-operated (GOGO); government-owned but contractor-operated (GOCO); owned and operated privately, with subsidies from the government to make the system competitive with other sources of power; or some mix of these.
As described in Chapter 3, in choosing among these variants, the nation should seek to minimize:
security risks (which argues for minimizing the number of sites involved and the amount of transportation of plutonium in forms vulner
See discussion in section "Economic Comparisons" in Chapter 6. Atomic Energy of Canada, Limited, the only vendor which has so far done a detailed study of possible use of the FMEF, concluded that adapting the facility to produce MOX for CANDU (Canadian deuterium-uranium) reactors would cost $118 million (AECL 1994).
able to theft, and maximizing the use of government sites, where requisite security either already exists or is simpler to provide);
costs and delays, both in construction and in gaining requisite licenses and approvals (which argues for making maximum use of existing or partly completed facilities, and again for use of government sites, where the licensing and public approval processes may be somewhat simpler); and
risks to environment, safety, and health (which argues for making use of the safest and best designed facilities likely to be available for the purpose, and for choosing the scale and types of nuclear materials processing so as to minimize risks and waste streams).
There are no options that perfectly meet these criteria. The following are a few of the most obvious specific candidates for this role:
Currently Operating, Utility-Owned Reactors: As noted earlier, three System-80 LWRs are operational at the Palo Verde site in Arizona, which could operate with full-MOX cores without modification, with license amendments.5 The same may be possible with a variety of other operational utility-owned LWRs in the United States. In one possible approach, for example, if the utility agreed to participate, the federal government would cover any additional costs in using government-furnished MOX fuel and would provide the necessary additional safeguards and security at the site, while the utility would otherwise continue to operate the reactors much as they are operated today. Additional financial incentives might be required to convince the utility to undertake the additional political and licensing burdens involved. As several sites have more than one reactor on-site, the handling of "fresh" plutonium and MOX fuel could be limited to two sites—one where the MOX fuel would be fabricated (presumably a site within the nuclear weapons complex) and the reactor site. Exploration of utility and public reactions to this concept is still in its early stages, though more than one utility has privately expressed interest to DOE.
The Washington Nuclear Project Reactors: Reactor 3 of the Washington Nuclear Project (WNP-3) in Washington state is a System-80 reactor, 75-percent complete, in the western part of the state, roughly 150 miles from the Hanford nuclear weapons complex reservation. WNP-2, complete and operating, is not a System-80, but may also be capable of handling a full core of MOX fuel without major modifications. It has the advantages of being complete, licensed, and located on the federal government's Hanford site, where the FMEF fuel fabrication facility is also located. The uncompleted WNP-1, like WNP-2, is not a System-80 but is located on the Hanford site. One or two of the three WNP reac-
tors could be acquired, completed (in the case of WNP-1 and WNP-3), and operated by the federal government (possibly in co-operation with a private entity) for the plutonium disposition mission. (For a discussion of the costs involved, see Chapter 6.) If the MOX fabrication capability at Hanford were used, this would have the significant advantage of confining all plutonium handling to two federal sites in the same state (or even a single large site, if only the WNP-1 and WNP-2 facilities on the Hanford reservation were used).
Two groups of private companies have put forward proposals for a government/private partnership to pursue this approach.6 If these options are to be preserved, action will have to be taken soon, as the WPPSS board has recently voted to cease maintaining WNP-1 and WNP-3 and to offer the components for sale.
Construction work on these reactors was halted some years ago. If these reactors were to be completed for MOX use, an intensive NRC review would be expected, which might result in requirements to upgrade some reactor systems to make them more comparable to the new reactors whose designs are now being reviewed by the NRC. Such modifications, if required, would add to the cost of completion.
DOE is also considering other options involving completion of partly-completed reactor facilities for the plutonium disposition mission, such as use of three Tennessee Valley Authority reactors on which construction has been
halted, but these would not have the advantage of co-locating fuel fabrication and reactor facilities at an existing government-owned nuclear site.
Acquisition of Existing LWRs: If the use of MOX fuel in utility-owned reactors or the WNP reactors faced insurmountable licensing or approval difficulties, another option is for the government to acquire existing reactors. There are several U.S. reactors that utilities may be willing to provide to the government, either because they were never completed or because their continued operation is becoming economically uncompetitive. 7 These could be acquired, modified if necessary for full-MOX operation, and used much as the WNP reactors might be. Such an acquisition would be unprecedented, however, and could raise a substantial set of procedural and institutional questions that could take time to resolve. No reactor operating license, for example, has ever been transferred from a utility to the federal government.
Principles for Institutional Arrangements: The specifics of such institutional arrangements require further study, but several basic principles suggest themselves:
the government should have a strong role to ensure that the approach fits with broader national policies relating to arms reduction and nonproliferation, that adequate security and safeguards are maintained, that any necessary openness to international inspection is maintained, and that appropriate ES&H standards are met;
the number of sites should be minimized to consolidate monitoring and safeguards functions and reduce the risks of plutonium theft;
if the sites were federal facilities (either already owned by the government or acquired for this purpose), it could ease the task of gaining the necessary approvals and licenses and maintaining the security and international transparency mentioned above;
any increase in government competition with private electricity generation should be minimized to the extent possible; and
if private investment can genuinely reduce government costs and upfront federal capital investments, it should be encouraged. But assessments of such possibilities must include realistic appraisals of all likely future costs and revenues, including the financial risks of government commitments to future subsidies or operations.
Approvals and Licenses
In addition to fuel fabrication, approvals and licenses will be important pacing factors for any use of plutonium fuels in LWRs in the United States. The United States initiated a licensing process for using MOX in LWRs in the 1970s (the Generic Environmental Statement on Mixed Oxide fuel, or GESMO) and obtained initial NRC staff approval, but this process was terminated when President Carter decided to end government support for the plutonium fuel cycle. While there do not appear to be fundamental obstacles to licensing U.S. reactors to handle MOX, the time to achieve the requisite fuel fabrication and reactor licenses is uncertain and might range from a few years to a decade, depending on the choice of reactors and sites as discussed above. Substantial public controversy could attend siting and construction of a plutonium fuel fabrication facility and use of plutonium fuel in U.S. reactors. There are important open questions concerning the licensing process for the various plutonium disposition facilities, and in particular whether oversight would be conducted by the NRC or the Defense Nuclear Facilities Safety Board (DNFSB). (See the appendix to Chapter 6 for more detail on the licensing issue.)
Public approval in the areas near the relevant facilities will also be a critical factor. Problems of public approval and licensing could be lessened somewhat if both the fuel fabrication facilities and the reactors handling MOX fuels were on federal sites. This is the main argument for building new reactors at existing DOE sites, rather than relying on existing civilian reactors. There is a good chance, however, that these problems could be addressed at some existing reactors.
Safeguards, Security, and Recoverability
The discussion to this point has focused primarily on feasibility and timing. Another important criterion identified in Chapter 3 is safeguards and security during the process. The panel believes that the United States and Russia should agree on a stringent standard of safeguards and security to be maintained throughout the disposition process. As described in earlier chapters, the panel recommends that safeguards and security for WPu be designed to come as close as practical to meeting the "stored weapons standard" throughout the disposition process.
An important issue is the risk of theft of materials during transport. Given the stringent security procedures and the low incidence of terrorism in the United States, this risk is likely to be greatly lower in the United States than it would be in Russia under current circumstances. (In both countries, however, it should be noted that thousands of assembled nuclear weapons, as well as other nuclear materials, are currently being transported each year as part of ongoing dismantlement programs and other activities.) The scale of transport required for disposition of excess WPu will depend to a great degree on the number of sites
involved—in particular, whether conversion of pits to oxides, fabrication of fuel, and the relevant reactors would all be at the same site or at several widely dispersed locations. The number of sites where this plutonium is handled, and the shipments of plutonium and their length, should be minimized to the extent possible to limit the risks of theft.
Once the plutonium is in the form of bulk oxide, rather than individually packaged pits, precise accounting to detect any diversion will become considerably more difficult. This will be a particular problem at the fuel fabrication facility, where the accounting system will need to have the capability for timely detection of diversion or theft of even a very small percentage of the facility's throughput. The International Atomic Energy Agency (IAEA) and EURATOM (the European Community's nuclear agency) (with assistance from the Los Alamos National Laboratory) have been working for years to develop new techniques for safeguarding such large plutonium bulk-handling facilities, as similar large facilities for civilian plutonium processing are scheduled to open soon in Europe and Japan. Nevertheless, some of these techniques are still in development, and it is doubtful that material accounting alone will be able to guarantee that diversion of enough plutonium to make a bomb could be detected within days. It will probably not be possible to achieve the “stored weapons standard" of accounting (described in Chapter 3, "Criteria for Choice") when dealing with complex, multistage processing of plutonium in bulk form. Therefore in addition to stringent material accounting, there should be extensive containment, surveillance, and security measures to ensure that no plutonium leaves the site without authorization.
Plutonium imbedded in the MOX spent fuel would be roughly as difficult to recover as the much larger and growing quantity of plutonium in commercial spent fuel worldwide, and this option therefore meets the "spent fuel standard." The MOX spent fuel, however, would have higher plutonium concentrations, so that a smaller total amount of MOX spent fuel would have to be processed to acquire enough material for a weapon.
As described in Chapter 3, in the case of reactors that would operate regardless of the mission of WPu disposition, the cost of using excess WPu in existing LWRs should be reckoned as the net additional cost compared to the cost of producing the same electricity by the means these reactors would otherwise use-that is, with LEU. As described in detail in Chapter 6, the required subsidy for using MOX fabricated from WPu rather than LEU in existing LWRs is likely to range from several hundred million to more than a billion dollars. If reactors had to be built, completed, or modified, or if the differences between LEU and MOX spent fuel involved higher disposal costs for MOX, these expenses would have to be added to this figure. In the case of the reactors that would not other-
wise operate, one must compare the cost of generating electricity in this way to the cost of generating it by the means that would otherwise be used—in most cases in the United States, either coal or combined-cycle natural gas. In effect, in these cases a twofold subsidy would have to be paid: the additional cost of using nuclear reactors rather than the least-cost means of electricity generation, and the net additional cost of using plutonium-based fuels rather than LEU in those reactors.
Environment, Safety, and Health
With appropriate modifications, the panel believes that it should be possible to operate U.S. LWRs with full-MOX cores while meeting the same safety standards that pertain to LEU fuel. Detailed safety analyses will be required, however. The plutonium processing necessary for this option (pit conversion and fuel fabrication) would inevitably result in wastes, risks of accident, and worker hazards. Careful design and the application of sufficient resources, however, should enable these facilities to comply with current regulatory standards. Commercial MOX operations are now routine in Europe and were carried out in the 1970s in the United States by several companies, but have not yet been undertaken in the current U.S. regulatory environment.
The spent fuel resulting from this option would be similar in most respects to ordinary LEU spent fuel, but there are differences.8 MOX spent fuel will contain more plutonium than typical spent fuel (raising potentially greater criticality concerns after eventual emplacement in a geologic repository) and will emit more heat for a longer time (which has an impact on the repository volume needed to hold a given number of fuel assemblies). The possibility that the somewhat different chemistry of the MOX spent fuel would affect long-term rates of release of radioactive materials in the repository would also have to be examined. This different spent fuel would have to be separately licensed as an acceptable waste form for geologic disposal, meaning additional costs and potentially additional delays. If the repository criticality issues are effectively addressed, however, it should be possible to store and dispose of MOX spent fuel as safely as LEU spent fuel. If the reactors used for this purpose would have operated with LEU in any case, the total amount of spent fuel to be disposed of in a geologic repository would not be increased as a result of plutonium disposition; even if reactors were operated specifically for plutonium disposition, the total amount of added spent fuel would be a small fraction of the planned capacity of the first U.S. repository. (See Chapter 6 for a more detailed discussion.)
As noted in Chapter 3, policy-makers considering plutonium disposition options should be aware that use of U.S. WPu in U.S. LWRs could be perceived by some as a significant change in U.S. policy, which has been not to pursue a plutonium fuel cycle, in part because of proliferation concerns. Such a perceived shift could have an impact on decisions on civil plutonium policies in Europe, Japan, and elsewhere. If the U.S. government wished to limit this effect and maintain its current policy of seeking to minimize worldwide stockpiles of separated, weapon-usable fissile material, it could make it clear that reducing the security dangers posed by existing stocks of excess WPu by using this material as reactor fuel, once-through, was perfectly consistent with such a policy. We note that since the WPu is already separated, the choice of a reactor option would not necessarily reopen the contentious question of reprocessing in the United States. All of the "elimination" options would require reprocessing and recycle, which could raise safeguards and security concerns, and would raise contentious political issues in the United States; such an approach would require a modification of current administration policy, which is not to reprocess plutonium for civilian energy generation.9 (See discussion in Chapter 6, "General Considerations.")
The "Spiking" Option
Another option making use of LWRs would be to "spike" the fuel by brief irradiation, processing the plutonium more rapidly, but creating a much less substantial radiation barrier. Such spiking would require a larger fuel fabrication facility (implying a higher capital cost) and more frequent reactor shutdowns for refueling (implying a lower capacity factor and more lost revenue).10 Expanded fuel storage capacities at the reactor sites would also be required to handle the fuel between the time when it was spiked and when it was recycled into the reactor to finish burning it to "spent" fuel. Hence, the costs of the spent fuel option would increase significantly if "spiking" were used as a first step. In the panel's judgment, the security for the material that could be gained by this more rapid but less extensive irradiation could be achieved more simply by providing appropriate security at the plutonium storage site; given its substantial costs, the spiking step on the path to the spent fuel option in LWRs is probably not worthwhile.
The "Elimination" Option
If excess WPu is used to fuel reactors on a once-through fuel cycle, the resulting spent fuel will add a small increment to the very large and growing amounts of plutonium in spent fuel around the world. Some analysts have advocated going farther, "beyond the spent fuel standard," seeking to minimize or eliminate this small increment to the existing spent fuel problem. Given the extra costs, delays, and risks involved in such an effort (described below), the panel does not recommend undertaking such a program for the narrow purpose of eliminating WPu. The broader issue of eliminating both WPu and the global stock of RPu can only be considered in the broader context of the future of nuclear power, which is beyond the scope of our analysis here. Nevertheless, we describe below some of the issues involved in such elimination approaches, in this case using LWRs.
Two possible definitions of "elimination" in reactors present themselves. On the one hand, one could seek to fission this particular plutonium as completely as possible. Plutonium would be loaded into reactors and burned; residual plutonium would be separated from the spent fuel, fabricated into new fuel, and burned in reactors again. Any reactor fuel cycle with reprocessing and recycle, and a conversion ratio less than unity, can obtain a high degree of elimination of an initial quantity of plutonium in this way.11 Complete elimination, however, is not possible, as some of the material will inevitably be lost to waste in the course of reprocessing and recycle.
Another, less demanding approach is to seek to eliminate the added increment of plutonium in spent fuel resulting from WPu disposition. If the WPu were used to fuel reactors that would otherwise have operated with LEU (and would have produced plutonium in the process), then when the excess WPu had been consumed to the point that the remainder was equal to the amount of plutonium whose production had been avoided by fueling these reactors with plutonium rather than LEU, the plutonium can be said to have been effectively eliminated. At that point, there would be no more plutonium in the world than
there would have been if the excess WPu had never existed, and any added security risk from the excess WPu would certainly have been eliminated. The same amount of electricity and the same mass of fission products would have been produced as would have been produced in the absence of the WPu. The only difference would be that some uranium would remain that would have been consumed had the excess WPu not been available as a substitute. In most cases, achieving such "effective elimination" would also require some degree of reprocessing and recycle. We consider "elimination" by these two definitions in the following paragraphs.12
Elimination by Fission
Flowsheets for a 1,000-MWe version of a System-80 reactor operating as a plutonium burner with plutonium recycle have been presented, assuming plutonium make-up is obtained by reprocessing uranium fuel discharged from LWRs.13 The system would be operated by fabricating excess WPu into fuel, burning it in the reactor, separating the residual plutonium from the spent fuel, fabricating that into fuel, and burning it in the reactor again. As the cycles continued, absorption of neutrons in the plutonium would cause the amount of higher isotopes of plutonium and higher actinides to increase, eventually reaching a steady state. This increase would substantially complicate reprocessing and fuel fabrication because of the higher radioactivity of the remaining actinides. It would also affect the reactivity of the remaining actinides, as the fission and absorption cross-sections of the various actinides vary considerably. (Indeed, no more than a couple of cycles have been accomplished in LWRs to date.) Assuming a capacity factor of 0.70, in the steady state such a reactor would consume approximately 0.50 tons of plutonium per year.
This does not mean that 50 tons of plutonium could be completely destroyed in 100 reactor-years. Such a linear destruction path would last only as long as there was enough plutonium remaining to keep the plutonium-burning reactors fully fueled. Once the amount of plutonium was reduced below the amount required to sustain a single reactor, additional fissile material (such as enriched uranium fuel in part of the reactor) would have to be added, to allow continued recycling for further destruction of the plutonium. The portion of the reactor devoted to MOX fuel would shrink with each irradiation cycle, until finally the desired degree of plutonium burnup was obtained. The destruction of
the residual plutonium in such a system is proportional to the product of the neutron flux and the amount of plutonium remaining in the core, so the destruction of the residual plutonium is like exponential decay. A finite time can be reported only if the extent of depletion is specified (see, e.g., Hebel et al. 1978, Choi and Pigford 1993).
A campaign to eliminate, for example, 99 percent of the WPu using LWRs would take many decades and would involve substantial costs and risks. No facilities designed for reprocessing LWR spent fuel are currently operational in the United States. Reopening of closed facilities or construction of new ones in the United States would be costly and time-consuming, and it would be expected to be the subject of considerable political controversy and intense regulatory scrutiny.14 Alternatively, reprocessing in other countries could be considered. Facilities capable of processing and fabricating multiple recycle plutonium (which is far more radioactive than WPu) would be needed, and technologies for that purpose would have to be demonstrated and licensed. The costs of repeated reprocessing and refabrication, and the costs of fuel development and licensing for multiple recycle fuel, would add a substantial increment (in the range of billions of dollars) to the subsidy required for the once-through plutonium disposition approach.
If the traditional PUREX process for separating plutonium from LWR fuel were used, such an elimination campaign could have the effect of increasing rather than decreasing net proliferation risk, as repeated and extensive handling of fully separated weapon-usable plutonium would be required. (The steady-state material that would eventually exist, however, with its high admixture of higher plutonium isotopes and other actinides, would be even less attractive for use in weapons than ordinary reactor-grade plutonium.) It is possible that some of the reprocessing approaches that have been suggested for other reactor types, which do not fully separate the plutonium, could be adapted to LWR fueling, but such approaches would have to be developed.
The net rate of plutonium destruction would be increased if additional plutonium was not being produced by absorption of neutrons in the uranium in the MOX fuel. Uranium-free "nonfertile" fuels for LWRs have been proposed for this purpose, containing plutonium, an inert material such as zirconium, and burble absorbers such as erbium (see ABB-CE 1993, pp. 111-51 ff.; and INEL 1993c). Use of such nonfertile fuels in LWRs has not been demonstrated on any substantial scale, and a substantial fuel development program would be required, involving significant cost and time. A variety of safety issues would require careful examination, and licensing of the new fuel could be expected to be difficult and time-consuming.
The advantage offered by nonfertile fuels for a plutonium elimination campaign may be less than is commonly thought. This is because simply increasing the plutonium concentration in ordinary (i.e., fertile) fuels would substantially reduce the amount of new plutonium produced for each unit of WPu burned or energy generated. For example, if the power level is to be kept constant, doubling the plutonium concentration (for example, from 3.5 to 7 percent of heavy metal atoms in MOX fuel for an LWR) requires, for a constant power, the same number of total fission per second.15 The rate of production of fission-spectrum neutrons is the same (for any given nominal reactor power), while the flux of fission-inducing neutrons is reduced, in order to keep the same integral of plutonium fission cross-section multiplied by flux. While the near-thermal ("epithermal") neutron flux would not be reduced by as much as a factor of two by such a doubling of plutonium concentration, it would surely be reduced, and the production of Pu-239 (via neutron capture by uranium-238 [U-238], followed by two beta decays) will be less for a given reactor power than with the lower WPu loading. This effect is slightly reinforced by the displacement of some fertile U-238 by plutonium, in consequence of the higher plutonium loading. Higher initial plutonium loadings would reduce the rate of new production still further. Since, in addition, even nonfertile fuels cannot burn their plutonium content down to zero (because at low enough concentrations of plutonium in relation to neutron-absorbing fission products, a chain reaction can no longer be sustained), it seems unlikely that the development of such fuels for reactors not already designed to use them could provide an advantage big enough to justify the required level of effort.
Any "effective elimination" campaign would use a reactor and fuel that would destroy a certain number of kilograms of plutonium per year. If we scale to 1,200 MWe, a comparable LEU-fueled LWR produces 253 kg of plutonium per year (see Table 6-1). Thus, for each year that a plutonium-burning reactor operates instead of a comparable LEU-fueled LWR, the amount of plutonium that otherwise would have existed in the world is reduced by the net plutonium consumption of the reactor, plus 253 kg. To "effectively eliminate" 50 tons of excess WPu would require running enough plutonium reactors long enough for this annual sum to add up to 50 tons.
To accomplish this objective using one 1,200-MWe LWR with a 100-percent MOX core would require over five decades of reactor operations, with two to three phases of separating the plutonium from the spent fuel and recy-
cling it (see Garwin 1995). For an LWR burning nonfertile fuel, the equivalent time is nearly four decades. Counting only current market prices for reprocessing and MOX fuel fabrication (not the substantial costs of development and licensing, building a new reprocessing facility if such were done in the United States, or the increased reprocessing and fabrication costs likely to be associated with multiple recycle material), "effective elimination" with LWRs using MOX would cost some $2.3 billion more than the once-through spent fuel option with similar reactors. (The times and costs for elimination by fission, as described above, would be substantially higher, as more cycles would be necessary.) All of the remarks above concerning required development, licensing and political issues, and security concerns would apply to this case as well.
RUSSIAN PLUTONIUM IN CURRENT-GENERATION RUSSIAN THERMAL REACTORS
Description of Technology and Status
The major differences in the case of using Russian LWRs to process Russian excess WPu include much higher security risks in the disposition process because of the current economic and political upheavals in the former Soviet Union; much lower availability of funds to finance the process; a smaller existing infrastructure of safe reactors; less experience in LWR MOX operations; and different economic conditions, plutonium fuel policies, and licensing procedures.
Russia has two major classes of thermal reactors, the VVER series LWRs and the RBMK graphite-moderated reactors. The VVER reactors use a conceptual approach similar to that of U.S. PWRs. Thus, much of the discussion in the first section of this chapter of the capability of existing LWRs for disposition of excess WPu is applicable to the existing VVER reactors and does not require repetition here.
VVERs have been built at two nominal power outputs: 400 and 950 MWe. Twenty-five VVERs are currently operable in the former Soviet Union (Nuclear News 1993). There are six nominally 400-MWe units in Russia and two in Ukraine. These reactors do not have containments, a major difference in safety from international standards. The early models (VVER 440-230) were not designed to withstand major earthquakes or the level of cooling water losses which Western reactors are designed to survive, have less redundancy in their safety systems, lack emergency operating procedures and training simulators to assist operation in responding to upset conditions, and otherwise fall far short of internationally accepted safety standards, such as those of the IAEA (IAEA 1992a). As a result, some of the VVER 440-230s have been shut down (in Russia and Armenia and also in eastern Germany). The later models (VVER 440-213) have incorporated some improved safety features, but still lack many key safety sys-
tems and other features and remain far from internationally accepted standards (IAEA 1992b). All the VVER-440 designs have more thermal operating margin than Western reactors, however, which has contributed to a good availability record. Since the operable VVER-440 units in the former Soviet Union have a small capacity, are old, and are substandard from a safety standpoint, the panel does not believe they should be used for the WPu disposition mission.
Seven nominally 950-MWe VVERs (model VVER 1000-320, usually referred to simply as the VVER-1000) are currently operable in Russia and ten in Ukraine. Twelve more are under construction in Russia and six in Ukraine, though construction is continuing on only a few of these. These larger VVERs have been provided with containments and have moved closer to Western designs both in increased safety and reduced thermal operating margins. The VVER-1000 reactors do not currently meet international safety standards, principally because they require thorough analysis and upgrading of (1) their instrumentation and control systems, (2) their use of redundancy to enhance safety margins, and (3) their emergency operating procedures and operator-training regimes. Nevertheless, the consensus of foreign experts is that, with the upgrades that are expected to be implemented over the next several years, they will be adequately safe. In any event, the Russian and Ukrainian governments are expected to operate these reactors for the long term.
An international program is underway to improve the safety of Soviet-designed reactors, but because the safety of the VVER-1000 reactors is so much better than that of the RBMKs and VVER-440s, the principal international safety effort today is directed at resolving the safety problems with these latter reactors except for upgrading emergency operating procedures and training; little work has been done on improving the safety of VVER-1000 reactors in the former Soviet Union, although there are plans for upgrading the plant systems in the future (EBRD 1993).
The capability of VVER-1000s to process WPu in the form of MOX fuel is likely to be similar to that of current-generation U.S. LWRs. None of the VVERs, however, have yet operated with MOX fuel, and the capability of each operating reactor in this regard is yet to be confirmed. Representatives of the Russian Ministry of Atomic Energy (MINATOM) confirm that studies of MOX use in VVER-1000s are just beginning. Preliminary Russian analyses have concluded that one-third MOX cores could be accommodated in VVER-1000s with little change in the reactors. Analyses of full-MOX cores are just beginning; as in other LWRs, use of plutonium fuel would reduce control-rod worth and the delayed-neutron fraction, which could require some modifications to ensure sufficient control capability (Levina et al. 1994, Novikov et al. 1994). More analysis is required to determine what measures would be necessary to ensure safety in an all-MOX VVER-1000, and in particular whether the VVER-1000s also have the specific design features that are reported to give many U.S.-
designed reactors enough control margin to operate with full-MOX cores of WPu without substantial modification.
The panel's general conclusion, which requires further analysis to confirm but is based on our knowledge of Western PWRs of similar core design, is that with core-reactivity-control backfits as necessary, VVER-1000s could be operated with full-MOX cores of WPu without eroding existing safety margins. If it is determined that significant reactor modifications would be required in the VVER-1000 case, it should be noted that each operational VVER-1000 is scheduled to be shut down for roughly one year for safety improvements under the ongoing program of international safety assistance. With enough lead-time for proper design and preparation, modifications that might be necessary to handle full-MOX cores could be made during this period, without substantially extending the length of the shutdown. Alternatively, VVER-1000s scheduled for completion in the near future could be modified for this purpose as they are completed.
Use of MOX in LWRs is not currently the Russian government's favored approach to dealing with excess WPu. Because of the delays in commercializing fast-breeder reactors that would consume plutonium separated by reprocessing, all other major reprocessing countries except Britain have decided to use plutonium as MOX in LWRs to limit the buildup of large stores of separated plutonium. Russia has not yet taken this route, preferring to save both military and civilian separated plutonium for eventual use in breeder reactors (see section "Advanced Liquid-Metal Reactors" below). Russia already has some 30 tons of civilian separated plutonium in storage, and more is building up every year, in addition to the excess military plutonium resulting from arms reductions.
Some representatives of MINATOM, in their advocacy of using WPu in fast reactors, have suggested that because Russia does not yet have experience with MOX in LWRs, the fast-reactor option could in fact be accomplished more quickly than the LWR MOX option.16 There are increasing indications, however, that MINATOM officials are beginning to look more favorably on the option of using plutonium fuel in LWRs that exist today, to deal with the increasing plutonium surplus in the near term.17 While Russia does not have direct experience of LWR MOX use, it appears that Soviet-designed LWRs, like their
foreign counterparts, could be prepared for MOX use on the basis of experience in other countries. And, as described below, under current economic conditions it is likely to be quite some time before sufficient funds to build the fast reactors MINATOM envisions become available.
The former Soviet Union also has RBMK reactors. Although their reactor coolant is light-water, RBMK reactors are substantially different from the LWRs discussed above and are not classified as such. The RBMK reactors are of a design unique among the world's present commercial electric-generating nuclear plants. The core consists of a large block of graphite moderator interspersed with pressure tubes containing low-enriched UO2 fuel, through which light-water is pumped to remove the heat from the rods. The reactor can be refueled continually while at power. The design is conceptually similar to U.S. plutonium production reactors and resembles the Hanford N Reactor, which was the only U.S. plutonium producer that also generated electricity. The N Reactor has been shut down permanently.
RBMKs have been built at two nominal power outputs: 925 and 1,380 MWe. Fifteen RBMKs are currently operable. Thirteen of them are nominally 925-MWe units—eleven in Russia and two in Ukraine (Chernobyl). An additional 925-MWe unit is under construction in Russia. There are two 1,380-MWe units, both located in Lithuania. The RBMK designs are further below international safety standards than the VVER designs. Reactor containment is not provided, there is limited protection from loss-of-coolant accidents, and the void coefficients of reactivity lead to core instability at low power levels. It was this type of reactor that suffered the disastrous accident at Chernobyl.
This panel has not evaluated the capability of the RBMKs to process WPu, although the reactor probably can operate with partial plutonium fuel loads with suitable modification for reactivity control. Because of the accident at Chernobyl, there has been substantial pressure, from within and without the former Soviet Union, to shut all the RBMKs down. The governments of Russia, Ukraine, and Lithuania have stated that the ensuing power shortages would threaten public safety more than continued operation of the plants, however, and have declared their intention to keep operating most of them. Debates on this subject continue. Efforts have been initiated by Western countries to assist Russia, Ukraine, and Lithuania in improving the safety of the RBMK and VVER 440-230. It is generally acknowledged, however, that it will not be practical to reach a level of safety conforming to international standards, and therefore, there is a desire to shut these plants down as soon as sufficient alternate capacity can be made available. The panel does not believe that the RBMKs should be considered for the plutonium disposition mission.
Reactor Throughput, Once-Through Fuel Cycle
If full-MOX cores proved acceptably safe, with enrichments of perhaps 5 percent plutonium in the fuel, two VVER-1000 reactors could transform 50 tons of WPu into spent fuel in 30 years. As there are four operational VVER-1000s at the Balakovo site, for example, all the reactor operations for the disposition mission could be carried out at a single site. If, on the other hand, these reactors were limited to one-third MOX fuel, at a relatively low enrichment of 2.5 percent, nine reactors would be required to accomplish the same task. As there are only seven operational VVER-1000s in Russia, either completion of additional reactors or use of some of the 10 operational VVER-1000 reactors in Ukraine would then be required. The Zaporozhe site in Ukraine has five plants in operation and one partly constructed; Zaporozhe and Balakovo together could handle an entire one-third MOX campaign at two sites. Use of weapons-grade plutonium in reactors in Ukraine would involve some important political issues, however, given the changing and uncertain relationship between Ukraine and Russia concerning both Ukraine's shipment of nuclear weapons back to Russia and Russia's provision of nuclear fuel for Ukraine's reactors.
As in the United States, the time at which disposition of excess WPu could begin would be paced by the availability of a MOX fuel fabrication facility. While Russia has laboratory-scale MOX fabrication facilities, no production facility with the required capabilities is currently operational.
A MOX fabrication facility with an intended capacity of about 100 MTHM/yr—enough to feed four VVER-1000s using full-MOX cores (processing as many tons of plutonium annually as the percentage in the fuel)— is reportedly roughly 50 percent complete at the Chelyabinsk-65 site. This facility was intended to produce MOX fuel for the planned BN-800 series liquid-metal breeder reactors. Construction on both this plant and the first of the BN-800s has been halted since well before the collapse of the Soviet Union due to lack of funds. Completing the plant and modifying it to produce LWR rather than liquid-metal reactor fuel would require several years at a cost in the range of hundreds of millions of dollars. The standards of safeguards, security, and ES&H this plant was designed to meet—or could practicably be modified to meet—are unknown.
Alternatively, a new MOX fabrication facility could be constructed, dedicated to the excess WPu disposition mission. The German company Siemens has proposed using disarmament assistance to build a replica of the Siemens fabrication facility already built at Hanau, which has a design capacity of 120 MTHM/yr. Siemens estimates the cost of building such a facility in Russia at roughly half a billion dollars, and believes construction could be accomplished in a relatively short period (Schmeidel 1992, cited in Berkhout et al. 1993).
Similarly, the French state-owned company COGEMA has expressed interest in participating in providing MOX fabrication capability for disposition of Russian WPu.
Reactor and Institutional Options
The public versus private issues in Russia are somewhat simpler than those in the United States, since MINATOM runs both the nuclear weapons complex and the civilian nuclear reactor industry. But because of the severe economic crisis in the former Soviet Union, U.S. or international financial assistance may well be required if long-term disposition of excess WPu in Russia is to be accomplished in the foreseeable future. Just as private investment might help reduce up-front capital costs in the United States, private investment or loans from international financial institutions such as the World Bank or the European Bank for Reconstruction and Development might help finance the operation in Russia, reducing the "line-item" costs that would have to be borne by any single government (see discussion of costs below). These institutions are already considering helping Russia to complete the VVER-1000 reactors under construction and to facilitate the shutdown of older unsafe reactors.
Approvals and Licenses
The political and institutional climate for plutonium use in Russia differs from that in the United States. In Russia, the government and the nuclear industry (controlled by MINATOM) are committed to a closed fuel cycle, including plutonium fuels, emphasizing fast-breeder reactors. MINATOM wishes to save the excess WPu for eventual use as startup fuel for future breeder reactors. Some others indicate a desire to sell the excess plutonium. Virtually all Russian government officials maintain that WPu has value that must be exploited.
According to news and other reports, however, the Russian public—after decades of government secrecy and the Chernobyl disaster—has become increasingly wary of all things nuclear and distrustful of all government assurances about environment and safety. Public resistance to plutonium use may therefore be significant. The regional and local authorities in Tomsk (a major production site for WPu), for example, have gathered sufficient strength in opposing the siting of a WPu storage facility there to call into question the viability of the site. The regulatory agency empowered (in principle) to regulate reactor siting and licenses, GOSATOMNADZOR, is seeking to define its role in the new Russia, and its future powers and attitudes toward plutonium use are uncertain.
Thus, the time required to gain the required licenses and approvals in Russia is more uncertain than in the U.S. case, and could ultimately prove to be either longer or shorter.
Safeguards, Security, and Recoverability
The risks of theft in transporting and processing plutonium in Russia under present circumstances appear high. Indeed, some analysts have argued that continued storage of the plutonium under high security until the Russian political and economic situation had stabilized would pose fewer risks than the processing and transport involved in the MOX option.
There are some important mitigating factors, however. As with the Hanford facility in the United States, the unfinished MOX fabrication facility in Russia is at a major nuclear weapons site. In addition, as noted above, the four VVER-1000s at the Balakovo site in themselves have sufficient capacity to carry out the WPu disposition mission if full-MOX cores were used. Thus, it might be possible to accomplish all processing of bulk plutonium at a single existing nuclear weapons complex site and all reactor use of plutonium fuel at one additional site. As in the U.S. case, all of the disposition steps should be subject to a stringent agreed system of safeguards and security.
This option meets the "spent fuel standard," though there would be significantly higher plutonium concentrations in the spent fuel than in typical commercial spent fuel.
Russian costs are uncertain, and no detailed analysis is possible with the information available. It is clear, however, that Russia has an overcapacity of low-cost LEU available for fueling its thermal reactors, which it is trying to market in the West to earn hard currency. It is also clear that significant up-front capital would be required to provide requisite plutonium fuel fabrication capability and to modify reactors to handle full-MOX cores. Therefore, substituting WPu for uranium in Russian LWRs would require a subsidy, probably in the range of hundreds of millions of dollars. The up-front capital investment (offset in part by later revenue) would be larger.
Because both capital and labor costs are substantially lower in Russia than in Western countries, however, it is possible that MOX produced in Russia would be competitive with LEU produced in the West. Contracts for sale of such MOX to Western utilities could serve as the basis for loans to provide the capital to build and operate a MOX fabrication plant. In this case, little if any explicit government subsidy might be needed to accomplish disposition of Russian WPu.
Environment, Safety, and Health
To a large extent, the ES&H impact of plutonium disposition in Russian reactors would depend on the resources applied to mitigate these impacts, as well as on the standards set. Standards for ES&H protection in the former Soviet
Union were low, and the resulting devastating environmental legacy is now becoming clear. New ES&H policies are evolving in Russia, with uncertain prospects. As noted above, while further study is required, it appears that the VVER-1000 reactors could be modified to handle full-MOX cores without decreasing their safety. The panel believes that with sufficient resources, MOX fabrication operations could be conducted in Russia (as in other countries) while meeting stringent standards of ES&H protection.
As with other options, policy-makers will have to consider how this approach fits with overall fuel-cycle policies. Assistance for using MOX in Russian reactors would provide a boost to the plutonium fuel cycle in Russia. There might also be some political impact in other countries whose civil plutonium programs are controversial (see discussion in Chapter 6, "General Considerations").
In addition to excess WPu, Russia has some 30 tons of separated civilian plutonium waiting to be fabricated into fuel, and more is being separated each year. Presumably an operational MOX fabrication facility in Russia, once available, would be used for more than just the excess WPu. Some Russian officials and European analysts have suggested that Russia should fabricate the civilian plutonium into fuel before beginning use of WPu, since the civil plutonium builds up dangerous radioactivity more quickly. Thus disarmament assistance for construction of a MOX facility might in effect sponsor civilian plutonium use in Russia—and commercial competition for MOX fabricators in Europe.
The Spiking Option
The issues facing the spiking option in the Russian case are similar to those in the U.S. case, described above. The reduction in reactor capacity factor resulting from more frequent refueling would have substantially greater impact in the former Soviet Union, however, because of the lack of electricity capacity margin in Russia and Ukraine. Reduction in overall electricity capacity there would not only create economic difficulties but would increase the pressure to extend operations of the RBMKs and VVER-440s to make up for the capacity loss of the VVER-1000s, conflicting with the goal of shutting down these less safe facilities.
The Elimination Option
To implement the elimination option in the former Soviet Union would require the utilization of reprocessing facilities as well as the power-reactor and MOX fuel fabrication facilities discussed above.
Reprocessing facilities exist in Russia for the extraction of plutonium from military production reactors-the source of the WPu which now requires disposition. In addition, there is a large civilian reprocessing plant, known as "Mayak," at the Chelyabinsk site, and consideration is being given to construction of a very large civilian reprocessing plant at the Krasnoyarsk site. Presumably, these facilities could be utilized, as in the case of similar facilities in the United States, to reprocess MOX fuel assemblies in the elimination option. The "head-ends" of the military reprocessing facilities would have to be modified to handle the MOX fuel from the power reactors. In light of the history of operations of these facilities in Russia, however, it is questionable whether they meet either adequate environment and safety standards or safeguards standards appropriate to this mission. Thus, the destruction option would probably require the construction of new reprocessing facilities, as might also be the case in the United States.
CURRENT-GENERATION CANDU REACTORS
Description of Technology and Status
In Canada, commercial nuclear power is obtained from CANDU (Canadian deuterium-uranium) reactors that use uranium fuel and pressurized heavy-water coolant confined to a system of pressure tubes in a calandria structure. The heavy-water moderator at low pressure is in the intervening spaces between the pressure tubes.18 The low neutron absorption by heavy water allows natural uranium to be used as fuel. Even so, only limited burnup of the natural uranium fuel is possible, limited by the low reactivity of natural uranium. This necessitates frequent replacement of irradiated fuel, accomplished by a pair of refueling machines that operate continuously while the reactor is at full power. Each fuel assembly is a cartridge of Zircalloy-clad rods of uranium dioxide, about 60 centimeters long. Each pressure tube contains several fuel cartridges. Discharged fuel is stored, to be later emplaced in a geologic repository. Twenty-two CANDU reactors are operating at high capacity factors and low fuel costs, at outputs varying from 515 to 881 MWe (Nuclear News 1993). Use of MOX in CANDU reactors is an option for WPu disposition. While this option appears technically and economically feasible (for either U.S. or Russian excess WPu), major political questions remain open. The panel notes that for this option vir-
tually all of the information made available to the panel was provided by the vendor, and had not yet been reviewed by DOE or other organizations.
Atomic Energy of Canada, Limited (AECL), the CANDU vendor, reports more than 25 years of experience with experimental irradiation testing of MOX fuels, with plutonium loadings between 0.5 and 3.0 percent. An experimental MOX fabrication facility was installed at Chalk River in 1975, and operated until placed on standby in 1987; this facility is now being reopened, and a resumption of MOX fabrication is expected in late 1995. More than 350 MOX fuel elements have been irradiated, in a zero-power test reactor, two research reactors at Chalk River, and in a prototype CANDU reactor. In the 1970s, six 19-element bundles containing 3 percent plutonium were irradiated to maximum burnups of 18,000 MWd/MTHM. One bundle remained in-reactor for 14 years, resulting in an outer-element burnup of 46,000 MWd/MTHM. In the early 1980s, six 37-element fuel bundles (similar to the bundles now proposed for plutonium disposition, but containing only 0.5 percent plutonium) were irradiated to burnups in excess of 21,000 MWd/MTHM. Thus the existing MOX irradiation experience includes the plutonium loadings and burnups relevant for plutonium disposition, with performance described as comparable to that of uranium fuel. Issues identified for further study are the effects on performance of the metal-to-oxygen ratio in the fuel and of the particle size (AECL 1994, pp. 2-57 to 2-59; Boczar et al. 1994).
The CANDU fuel element contains UO2 pellets in Zircalloy cladding. The fuel design has evolved toward the current use of thin cladding, with a minimum thickness of 0.38 mm. This thin cladding improves neutron economy (an important factor in a reactor fueled by natural uranium), but the lifetime of the cladding may limit the burnups that could be achieved with fuels enriched with plutonium.
The manufacturer reports current average discharge burnups of approximately 8,300 MWd/MTHM of uranium, with maximum local burnups in the range of 15,000 MWd/MTHM. Although both the MOX experience described above and the occasional occurrence of maximum bundle average burnups of over 25,000 MWd/MTHM during CANDU operation suggest that CANDU fuel can withstand high burnups (no performance limits resulting from the high burnup operation had been identified at the time of a 1983 summary presentation on fuel performance) (Gacesa et al. 1983), representatives of the vendor indicated that they would not have high confidence in cladding performance at burnups significantly beyond the current nominal maximum local burnup without extensive further testing (Feinroth 1993). The vendor's reference case calls for an average burnup of 9,700 MWd/MTHM with MOX fuel if the current fuel design is used, or 17,700 MWd/MTHM if the new CANFLEX design now being tested is used (AECL 1994).
CANDU reactors appear to be capable, without physical modification, of handling 100-percent MOX cores. The vendor concludes that: "There should
not be any significant changes in the design safety margins of the CANDU MOX fueled reactor systems relative to the current natural uranium fueled operations" (AECL 1994, p. 8-1).
As in the U.S. LWR case, however, no CANDU reactors are currently licensed to use MOX fuel, and favorable regulatory review of the safety of their operation in this mode would be required. While the CANDU reactor design is in principle even more easily adaptable to full-core MOX operation than most LWRs, at the same time the technical uncertainties concerning MOX use in CANDUs must be considered somewhat larger than in the LWR case, given the lack of MOX operating experience in CANDU reactors. There are also considerable uncertainties concerning the economics of this option, as no one has ever produced CANDU MOX fuel before.
The vendor suggests that the four 825-MWe units located at the Bruce A site would be particularly suitable for WPu disposition. Ontario Hydro, the utility that owns these reactors, has expressed interest in studying the plan. The utility has taken part in an initial study funded by DOE, and it sponsored a public meeting near the plants in June 1994 to discuss the idea of using plutonium there (AECL 1994, pp. 3-4 to 3-14).
AECL has also pursued preliminary discussions with the Russian government regarding this concept. The two sides signed a document calling for joint study of the use of Russian plutonium in CANDUs, if a third party (presumably the U.S. government) would provide funding. It is conceivable that production of CANDU MOX in Russia, with that country's low labor and materials costs, would be cheap enough that it could be sold to Canada at a modest profit (Feinroth 1994). (As noted above, this is also possible in the LWR MOX case.) While new CANDU reactors could also be built in Russia to consume the plutonium there, this would be slower and more expensive than using existing reactors, and would not appear to have great advantages compared to constructing reactor types with which Russia is more familiar, such as light-water reactors.
Compared to the use of U.S. LWRs, the use of CANDU reactors would have both advantages and disadvantages, discussed in the following subsections.
Straightforward Adaptation to Full-MOX Core Operation. As noted, it appears that existing CANDU reactors, without physical modification, could operate with full-MOX cores. In normal CANDU operations with natural uranium fuel, more than half of the energy is provided by fissioning plutonium produced in the fuel as the reactor operates. As a result, adding plutonium to the initial fuel would represent a smaller change in the physics of the reactor core than in the case of LWRs. Moreover, the structure of the CANDU reactors allows plenty of space for added controls, and additional neutron absorbers could be dissolved in the heavy-water moderator used in the reactors. Thus the vendor
concludes that no physical modifications would be required to handle substantial quantities of plutonium in CANDU reactors. Several factors should be mentioned in this regard.
First, CANDU reactors offer greater flexibility for maintaining control-absorber worth with plutonium fuels. In a calandria-type heavy-water reactor most of the reactor-core volume is occupied by the relatively cool, low-pressure, heavy-water moderator. In the CANDU design the movable control absorbers are located in the moderator region. There is ample space to increase the size of each absorber, or increase the number of absorbers, as necessary to counteract the higher neutron absorption in plutonium fuel and to maintain control-reactivity worth. The CANDU design also provides for boron absorber dissolved in the moderator for shim control of reactivity. The well-thermalized neutron spectrum provides greater control worth for a given amount of boron than in a plutonium-fueled LWR. The movable and soluble absorbers are also more effective because they are located in the higher-flux region where thermal neutrons are formed. Thus, the existing CANDUs can provide sufficient reactivity control to handle full-MOX loading. The designers expect that the reactivity control would even be sufficient for a full loading of nonfertile plutonium fuel.
Second, the CANDU design features thermal decoupling between the moderator and the fuel. The heavy-water moderator is at low pressure and is thermally insulated from the hot fuel and coolant in the pressure tubes. Therefore, there is only a weak coupling between an assumed sudden increase in power and temperature of the fuel and the temperature of the moderator.19 Neutron temperature effects are, therefore, of less concern than in LWRs, where the moderator is also the coolant, whose temperature increases rapidly following a spurious increase in local fission rate. Also, in the CANDU, dissolved boron for shim reactivity control is in the moderator rather than in the coolant. In a pressurized-water reactor, with boron dissolved in the coolant, thermal expansion of the coolant accompanying an assumed sudden increase in local fission rate reduces the amount of boron near the fuel and tends to add to the reactivity.20 But there is no boron in the heavy-water coolant of a CANDU reactor, and relatively little reactivity is added by moderator heating because of the weak thermal coupling between fuel and moderator.
Third, prompt neutrons have a considerably longer lifetime in a CANDU reactor than in other reactors because of the relatively large volume of heavy water with weak absorption of thermal neutrons. This results in a much longer
time-constant for power changes following an accidental insertion of reactivity sufficient to initiate a prompt-critical excursion. Thermal shock is less likely, and more time is available for safety-system actuation.
Finally, the addition of dysprosium in the MOX fuel has a beneficial effect on the reactor's performance in a loss-of-coolant accident (LOCA). CANDU reactors fueled with natural uranium will become more reactive in the event of a LOCA. For this reason, CANDUs are equipped with two independent fast-shutdown systems. The plutonium-dysprosium fuel is designed to counter this positive reactivity coefficient, so that the reactor will become less reactive in the same scenario. Analysis by the vendor shows in the event of a large-break LOCA, power output in a uranium-fueled reactor could increase to over four times its nominal value, but would not increase at all in the MOX fueled system (AECL 1994, p. 2-101).
Simplified Fuel Fabrication. CANDU fuel is produced in smaller and simpler units than those typical of LWRs, potentially reducing the fabrication cost, which is a substantial fraction of the total cost of MOX use. The specifications that CANDU fuel must meet, in such areas as granularity, pellet shape, and the like, are less stringent than those required for LWR fuel. Experience with LWR fuel, however, suggests that the need for tight specifications increases with greater enrichment, so adding plutonium to CANDU fuel might require specification tolerances closer to those of typical LWR fuel.
Continuous Fueling Offers Additional “Spiking" or High Burnup Options. Operating with natural uranium fuel requires frequent replacement of fuel, so CANDU reactors use mechanisms for continuous refueling while at power. The refueling mechanisms require relatively short (60 cm) fuel-assembly cartridges. As irradiated fuel cartridges are removed from one end of a pressure tube, fresh cartridges are inserted at the other end. Within a given pressure tube the cartridges can be moved along the pressure tube at a rate proportional to the average neutron flux in the pressure tube. All discharge cartridges can be irradiated to the same burnup, not possible in LWRs refueled by periodic batch replacement.
The ability to refuel without shutdown offers additional options for the "spiking" option. While the "spiking" approach would still require added capital expenditures for a larger fuel fabrication facility, it would not decrease revenue as a result of reactor downtime for refueling. Moreover, the remotely operated refueling machines would reduce concerns regarding possible worker exposures to radiation in reloading "spiked" fuel to finish burning it to "spent" fuel.
The small cartridges and remotely operated refueling machines also make it possible to consider recycling discharged cartridges for further irradiation, providing a possible means of greater annihilation of plutonium than otherwise obtainable in once-through irradiation. Fresh first-cycle plutonium-fueled cartridges could supply neutrons to overcome the neutron-absorbing fission-
product poisons in nearby recycled cartridges. Irradiation of individual cartridges could proceed to the material limits of the fuel material, rather than being limited by reactivity considerations. Similar operation of a plutonium-burning LWR would be difficult because of the very long (approximately 5-meter) fuel assemblies and the need to shut down and depressurize for refueling.
High Neutron Economy Potentially Contributes to High Burnup on Once-Through Cycle. The high neutron economy of the CANDU makes it possible to irradiate fuel of a given fissile concentration to much greater burnup, and to a greater degree of annihilation in one irradiation cycle, than is possible in other thermal reactors, such as the LWR or high-temperature gas-cooled reactor (HTGR). A measure of the neutron economy is the conversion ratio, defined as the number of neutrons absorbed in fertile material per fissile atom destroyed. For reactors fueled with U-235 in U-238, the conversion ratio of the CANDU is 0.75, compared with about 0.6 for a LWR or an HTGR. High neutron economy would be important for plutonium-burning fuel cycles that sought high burnup during an irradiation cycle using fertile-free plutonium fuels. (It still may not be possible, however, to design fuels that can achieve the very high burnups claimed for HTGRs, described below.)
In designing thermal reactors for efficient utilization of plutonium as a fuel, it is important to seek a well-thermalized neutron spectrum to avoid the higher capture-to-fission ratio of Pu-239 that results when the thermal-neutron spectrum is shifted to higher energies. Calandria-type heavy-water reactors have characteristically lower-energy and better-moderated spectra of thermal neutrons than do LWRs, particularly with plutonium fueling. Although Pu-239 is consumed by both neutron capture and fission, the nonfission capture produces additional plutonium isotopes such as Pu-240, 241, and 242. If disposition of WPu emphasizes the destruction of all plutonium isotopes, nonfission capture of neutrons in Pu-239 is less productive in destroying plutonium than using those neutrons for more extended fission of Pu-239 during an irradiation cycle.
Uncertain Canadian Acceptance. The use of existing CANDUs would have to be approved by the Canadian government, the reactor operators (primarily the Ontario Hydro utility), and the relevant regulators (the Atomic Energy Control Board). AECL, the government-owned designers of CANDU, and Ontario Hydro both participated in the initial DOE-funded study of the concept, completed in mid-1994, though neither has made a firm political commitment to support it. The Canadian government has reportedly suggested to U.S. representatives that the two countries form an expert group to explore the idea. But further discussions between the U.S. and Canadian governments would be required before it could be determined whether this approach had enough political support to be a
practical option. Canada has previously avoided using either enriched uranium or plutonium fuels in CANDU reactors, and might reject this plutonium-use option as well. Yet Canada has also traditionally played an active role in disarmament; playing a central role in disposition of materials resulting from nuclear arms reductions might well be appealing enough to overcome the resistance to use of weapons materials. Canadian public acceptance is also an open question.
Large-Scale International Plutonium Transport. The distance over which plutonium would have to be transported to be burned in CANDU reactors would be greater than that in using U.S. LWRs, even if all the CANDU reactors involved were at a single site. The attendant controversies and risks of theft would be correspondingly larger. Possibly more important in political terms than the sheer distances is the need for the material to be shipped across international borders, to a non-nuclear-weapon state.
Lower Plutonium Loading Requires More Fuel Fabrication. As described in more detail below, the CANDU manufacturer suggests WPu loadings of only 1.5-2.7 percent. This would require larger quantities of fuel to be produced than in the case of LWRs, increasing costs and countering the advantage of simpler fuel fabrication described above.
Lower Radioactivity and Small Size of Bundles. Because of the relatively low burnup (even when enriched with plutonium) and small size of the CANDU MOX bundles, the gamma-radiation dose rates from them would be somewhat lower than those from LWR spent fuel of equal age.21 The spent CANDU MOX, however, would have substantially higher dose rates for several decades than the large quantities of much older LWR spent fuel that will exist at the time the CANDU MOX spent fuel would be discharged. The small size of the CANDU bundles would make them slightly simpler to steal.
Safeguards Issues of Online Refueling. Fuel can be removed from CANDU reactors at any time without shutdown of the reactor, and the fuel elements are substantially smaller and more portable than is the case for LWRs. Therefore, CANDUs require more intensive safeguarding than do LWRs. For fuel containing more plutonium, still more intensive safeguarding would be needed. Both CANDU reactors and the fresh MOX fuel in store at either an LWR or a CANDU, however, require continuous safeguarding in any case. In addition, the task of accounting for and securing complete fuel assemblies for either a CANDU or an LWR is substantially easier than that of accounting for bulk plutonium at a MOX fabrication plant. Therefore, the net additional security risks
The surface dose rate 10 years after discharge from a single bundle irradiated to 9,700 MWd/MTHM is about 5,500 rem/h (roentgen-equivalent-man per hour) (AECL 1994, p. C-19). compared to a surface dose rate of 18,000 rem/hr at the same time for a pressurized-water reactor fuel bundle irradiated to 40,000 MWd/MTHM (see Table 6-5). The dose rate also falls off more rapidly with distance for the CANDU fuel bundle, because of its more compact size.
of using CANDU reactors for this mission compared to using LWRs would be relatively small.
Reactor Throughput, Once-Through Fuel Cycle
AECL Technology (the U.S. subsidiary of AECL) describes two main fuel options for burning plutonium in CANDU reactors (AECL 1994). As a reference case, AECL chose the Bruce A reactors, which each have a unit core thermal power of 2,832 MWt, providing 825 MWe of gross electricity. Since the units also produce excess steam for local heating, the total gross electrical equivalent output is 904 MWe per unit. Four of these units are in operation at the Bruce A site. The reference case of natural uranium fueling (i.e., roughly 0.7 percent U-235) operates at a typical discharge burnup of 8,300 MWd/MTHM.
The first option for plutonium disposition, using fuel elements similar to those now used with natural uranium, would have an average plutonium loading of 1.5 percent. The innermost pin and the first ring of pins surrounding it (7 of the 37 pins in the bundle) would consist of 5-percent dysprosium-oxide burble absorber in depleted uranium. These absorber pins would compensate for the greater reactivity of the plutonium fuel compared to natural uranium; in a CANDU spectrum, the dysprosium absorption would be reduced at approximately the same rate as the plutonium reactivity, helping to flatten the reactivity of the fuel over its life. The next ring would have 2-percent plutonium oxide, and the outermost ring 1.2 percent, for a total of 0.23 kg in each bundle. Average discharge burnup would be 9,700 MWd/MTHM, but because the combination of MOX and dysprosium absorbers would help to flatten the power distribution through the core, peak burnup would be increased only slightly compared to the reference case (15,500 MWd/MTHM versus 15,000).
With this approach, each reactor would consume 1.05 tons of plutonium per year; two of the Bruce A reactors could dispose of 50 tons of WPu in 25 years. No hardware changes to the reactor system would be needed to operate within the existing safety envelope, according to the manufacturer.
The second concept considered by AECL Technology is to use a new fuel design still undergoing testing, known as CANFLEX, designed to achieve higher burnup. CANFLEX bundles contain 43 elements rather than 37. Demonstration irradiations of this design in power reactors are expected to begin in 1996 (AECL 1994, p. 2-77). In this case, AECL Technology again envisions the innermost seven pins being dysprosium oxide (6.0 percent in natural uranium), with the outer ring being 2.1-percent WPu, and the intermediate ring 3.6 percent, for a total of 0.38 kg per bundle-nearly twice the plutonium per bundle of the previous case.22 This much larger amount of plutonium, however, is com-
pensated directly by a much higher burnup (17,100 MWd/MTHM versus 9,700 MWd/MTHM), so that the total amount of plutonium consumed in each reactor each year is roughly the same, still requiring 50 reactor-years for disposition of 50 tons of WPu.
The primary potential advantage of this latter approach is not speeding the process, but drastically reducing the number of MOX fuel bundles that would have to be fabricated, thereby potentially reducing the cost of the operation. The disadvantage of this approach is that CANFLEX fuel is not yet licensed, and a more substantial delay before CANFLEX MOX fuel could be produced would be expected. The manufacturer, therefore, suggests beginning an initial program with MOX in standard CANDU bundles, switching to CANFLEX when it becomes available.
Like the United States, Canada has no MOX fuel fabrication capacity. Fabricating MOX fuel for CANDUs at the Hanford FMEF facility would be the most expeditious approach, with the same caveats as in the LWR case. The vendor has in fact examined fabrication of MOX fuel in the FMEF in considerable detail, and believes that large throughputs of CANDU MOX fuel (over 160 MTHM/yr) are possible, by taking advantage of additional floor space not used by the current MOX fabrication line in the facility (AECL 1994).
Approvals and Licenses
Gaining approval of the various Canadian institutions and the Canadian public would be a major hurdle for the CANDU option. Licensing reactor operations with plutonium would probably be a less difficult issue than securing agreement on the basic approach. Licensing procedures and standards for plutonium use in Canada, set by the Atomic Energy Control Board (AECB), are different from those used by the NRC. In general, the process in Canada relies more on co-operation between licensees and the board, and less on an adversarial process.
Safeguards, Security, and Recoverability
The safeguards concerns regarding fuel fabrication are similar for LWRs and CANDUs. Because of the need to transport plutonium over longer distances, transport risks would be somewhat greater for CANDUs, and because of
the reactor's online refueling capability and the portability of the fuel elements, the risks of theft or diversion of fabricated fuel from the reactor could be somewhat greater as well. Both of these risks could be reduced to very low levels with the application of sufficient resources.
This option would make the plutonium roughly as difficult to recover as the plutonium in commercial spent fuel. As mentioned above, however, the generally lower burnups in the CANDU case and the small size of the spent fuel cartridges mean that the radiation field from each unit of spent fuel is somewhat lower.
The cost of this option is difficult to estimate, as experience fabricating MOX fuel for CANDU-type reactors is very limited.23 On the one hand, an argument can be made that the subsidy required would be less than in the LWR MOX case, because: (1) the fuel is simpler and probably cheaper to fabricate; and (2) at least in the case of the advanced CANFLEX fuel, the MOX fuel would have a higher energy content (and hence a longer fuel life) than the natural uranium fuel CANDU reactors normally use, so that the increased per-kilogram cost of fabricating the MOX fuel would be compensated, in part, by the reduced amount of fuel that would have to be fabricated. On the other hand, the subsidy required might also be more than in the LWR case, as the amount of natural uranium CANDU fuel each kilogram of MOX would substitute for—whose cost would be subtracted from the MOX cost in calculating the subsidy required—would be more than $1,000 cheaper than the LEU LWR fuel a kilogram of MOX could substitute for.24
The vendor estimates, probably optimistically, that the U.S. FMEF fuel fabrication facility could be modified for an overnight cost of $118 million, and produce MOX with annual operating costs of about $64 million per year for the reference MOX fuel and 20 percent more for the CANFLEX fuel option. As described in detail in Chapter 6, this implies total costs comparable to the middle of the range for the options that employ currently operating LWRs.
Environment, Safety, and Health
Use of plutonium in CANDU reactors would raise the same general concerns as those described for LWRs. In this case, because of the higher burnup of
the MOX compared to the natural uranium fuel it would replace, the total mass of spent fuel requiring disposal would be somewhat reduced.
The potential perceived impact of this approach on fuel-cycle policies would be more complex than in the U.S. LWR case. On the one hand, by providing excess plutonium free of charge to another nation, the United States would be demonstrating that it saw no economic value in the material and was encouraging its use in reactors only as an arms-control measure. On the other hand, the United States would still be encouraging use of separated plutonium as reactor fuel on a scale wider than would otherwise be the case in a non-nuclear-weapons state (see discussion in Chapter 6, "General Considerations").
The Spiking Option
The "spiking" option is also a possibility for the CANDU case. For spiking, the refueling machines would be operated at their capacity of 22 fuel bundles per day. The discharge burnup in this case would be only 5,100 MWd/MTHM. The plutonium processing rate per reactor would therefore be nearly doubled. By the same token, however, twice as much fuel fabrication capability would be required, potentially incurring significant additional capital cost. Arrangements for storing the radioactive "spiked" fuel and then eventually reinserting it into the reactor for burning to the spent fuel standard would be required.
The Elimination Option
The CANDU system, with its high neutron economy and on-power refueling offering flexibility in fuel management, could offer an opportunity for achieving substantial burnups without reprocessing if new fuels were developed for that purpose. The manufacturer suggests that development of a nonfertile fuel could allow nearly 60-percent destruction without reprocessing or fuel shuffling; higher destruction fractions could be achieved by fuel shuffling, that is by moving the burned bundles to other locations for additional irradiation.25 Development of such a nonfertile fuel might be easier for the CANDU because of the lower fuel temperatures.
Because of the flexibility of the CANDU control system, as described earlier, it might be possible to control the large reactivity changes during an irradiation cycle using nonfertile fuel more easily than in an LWR. The higher neutron economy of the CANDU should make it possible to maintain criticality to a greater fraction of initial fissile atoms burned than in an LWR. This, together
with the ability to discharge fuel to a more uniform burnup, should result in a greater fractional burnup of initial plutonium in a given irradiation cycle than in an LWR.
The relatively small fuel bundles of the CANDU and the remote refueling machines invite the possibility of recycling discharged high burnup fuel for further irradiation and burnup. The recycled discharged fuel could be located near fresh make-up fuel to maintain reactivity and promote further burnup.
As with LWRs, more complete elimination could be pursued with reprocessing and recycle, either of MOX fuels or of nonfertile fuels. From a technical point of view, there are no reasons to exclude CANDU reactors if reprocessing and recycle to eliminate WPu are to be considered for any reactor concepts. Like LWRs, however, a reprocessing cycle for CANDU reactors as currently conceived would involve substantial handling and processing of separated, weapon-usable plutonium, raising some security risks. New reprocessing facilities would have to be provided.
POTENTIAL INVOLVEMENT OF WEST EUROPEAN AND JAPANESE FACILITIES
Description of Technology and Status
Under established civil plutonium fuel programs, commercial reactors in Europe and Japan plan to process well over 100 tons of civilian plutonium over the next decade. Plutonium storage and transport arrangements, fuel fabrication capabilities, and reactors licensed to handle plutonium for this task already exist or are planned.26 Technical feasibility is amply demonstrated.
One possibility for long-term disposition of excess WPu, therefore, is to substitute this weapons material for the civilian plutonium. 27 Pits would be processed to plutonium oxide in their country of origin and the resulting oxide shipped to Europe or Japan for fabrication and use.28 In particular, such an approach would enable disposition of Russian plutonium in the near term while
minimizing the safeguards and security concerns that would be involved in large-scale MOX fabrication in Russia during the current period of high risk of theft.
In this case, the initial processing and shipment step would be the only aspect of plutonium handling beyond that already planned—with the important caveat that all these facilities would now be using weapons-grade rather than reactor-grade plutonium. From the point of view of civilian nuclear-energy production, the WPu would be less radioactive (and therefore easier to fabricate) and have slightly higher energy content than the reactor-grade material it would replace.
What would happen to the displaced civilian plutonium? Three main possibilities exist: one is to expand MOX operations in these countries, involving more reactors and fabrication facilities than those currently planned, so as to process both the civilian and the excess WPu. This approach would, however, mean an expansion in global handling and use of separated plutonium (as would options involving MOX use in the United States and Russia). MOX fabrication capacity worldwide is in a period of rapid expansion, with facilities larger than any that now exist scheduled to come on line in the next few years in France, Britain, and later in Japan as well. These facilities are sized to provide adequate capacity to balance the plutonium to be separated by civilian reprocessing in the same countries. A large MOX plant is nearly complete at Hanau in Germany, but has not been allowed to open because of licensing difficulties arising from political opposition in the state of Hesse. Many of the German utilities have succeeded in acquiring MOX contracts elsewhere, and some have begun to cancel their post-2000 reprocessing contracts with the French state-owned company COGEMA and British Nuclear Fuels, Limited (BNFL). The Hanau plant's capacity could potentially be used to fabricate either U.S. or Russian WPu (or both) into MOX fuel, if a political arrangement allowing it to be opened for this purpose could be reached. This general type of approach could be particularly attractive if the owners of the plants with such excess capacity, having paid the capital cost of the plants with the MOX fabrication contracts already in hand, would be willing to contract for additional fabrication at the cost of operations. If capital cost were excluded, MOX could be fabricated from free WPu for less than the cost of LEU. Russian plutonium, for example, might be shipped to Germany and fabricated into MOX which would be sold to German utilities at the price of LEU; the difference between this price and the operations cost of producing the MOX could go to Russia to pay the costs of secure storage and processing the pits to oxide.
Another possibility would be to continue reprocessing and MOX use as planned, and store the separated reactor-grade plutonium displaced by the WPu. The net result would be to convert an excess stock of separated weapons-grade plutonium to an excess stock of separated reactor-grade plutonium of roughly equal size-a step the panel considers to be of too limited benefit to justify the
complications of the required international agreements and the risks of the required international transport.
The third possibility is to defer reprocessing until existing excess stocks of separated plutonium (both weapons-grade and reactor-grade) are consumed. Reprocessing plants would be kept in cold standby until then.29 This approach would consume both the projected surplus of WPu and the projected surplus of separated civilian plutonium, without predetermining fuel-cycle choices for the period after the current stocks of separated plutonium are consumed. Such an arrangement, however, would require complex international agreements altering a web of existing contracts and spent fuel management policies.
Reactor Throughput, Once-Through Cycle
If the necessary agreements could be reached expeditiously, this would be by far the most rapid reactor option, since the pacing steps of building new fabrication capacity and licensing the various facilities would be avoided; as noted, over 100 tons of plutonium are expected to be processed in this way over the next decade in any case, so it would be technically possible to process the entire stock of U.S. and Russian excess WPu over that period. Reaching the necessary agreements could involve extended and unpredictable delays, however, with no guarantee of ultimate success.
Approvals and Licenses
Gaining the international agreements necessary to ship tens of tons of weapons-grade material from Russia or the United States to Europe would be difficult, and the amount of time required is impossible to predict. The highly enriched uranium (HEU) purchase agreement already negotiated between Russia and the United States could provide a model—though it should be noted that as of late 1994, three years after such an arrangement was first proposed, none of the material had yet been delivered.
Gaining agreement to the option involving a deferral of existing international reprocessing contracts would be particularly problematic. France and Britain share much of the world market for commercial reprocessing, and they have just completed multi-billion-dollar investments in new facilities. Any proposal to defer reprocessing for an extended period would be seen as a threat to these businesses. Even substantial financial compensation might not be suffi-
cient to overcome such objections. A multinational negotiation would be required in a forum not yet defined.
If some relevant countries were interested in pursuing this option, but others were not, the substitution of WPu for civilian separated plutonium might be only partial. Britain, for example, might agree to defer operation of THORP (Thermal Oxides Reprocessing Plant) and fulfill its contracts with WPu instead, even if France continued with its reprocessing operations as planned.
If reprocessing were deferred for an extended period, more spent fuel storage would be required. From the point of view of utility owners of nuclear reactors in countries such as Germany and Japan, the opportunity to dispose of their spent fuel is one of the primary advantages of reprocessing. These utilities might be very reluctant to agree to an additional decade's worth of spent fuel simply building up at their reactor sites.30 It is also an open question whether the public in France and Britain would accept the idea of highly radioactive spent fuel continuing to be shipped from abroad to reprocessing sites in their countries for storage, with no reprocessing planned for years to come.
Convincing the Russian government to accept such an approach would almost certainly require financial compensation for the material, as in the case of the HEU deal. Even then, some opposition to sending large amounts of a key strategic material abroad could be expected.
The international controversy provoked by the recent shipment of 1.7 tons of reactor-grade plutonium oxide from France to Japan suggests the political difficulties that would be faced by the much larger shipments of weapons-grade plutonium required for the plutonium disposition mission. To displace civilian plutonium to be used in Europe with Russian excess WPu, however, would require only overland transportation, which is less controversial than recent well-publicized shipments by sea have been. The association with arms reduction should also help reduce public criticism. Shipment of large quantities of weapons-grade plutonium, rather than merely reactor-grade plutonium, to non-nuclear-weapons states such as Japan and Germany would almost certainly arouse controversy with neighboring states, however.
Safeguards and Security
Since in this case WPu would displace separated plutonium operations that would take place in any case, the net additional safeguards issues involved in this option are smaller than those in other cases. The net additional risks would come from the pit processing required for all options; the large-scale interna-
tional shipment of plutonium, central to this option; and the difference in proliferation risk involved in the shift from reactor-grade to weapons-grade plutonium. The key potential advantage of this approach, from a security point of view, is that disposition of Russian plutonium could be accomplished quickly, with most of the bulk handling of separated plutonium required occurring in other countries, where theft risks currently appear to be lower. The need for an agreed, international approach to safeguards and security is even more obvious here than it is in other cases. The risks involved in the large-scale international transport of plutonium required in this option are difficult to judge and depend on the resources applied to reducing them.
In this option, a variety of parties would probably demand financial compensation for the materials used, the use of facilities, or the disruption of previous plans. Russia would probably insist on financial compensation for plutonium used abroad in this way, making it effectively a plutonium purchase arrangement similar to the HEU deal. MOX fabricators could be expected to charge at least the cost of operations, and possibly the full commercial rate for MOX fabrication. In the variant involving a deferral of reprocessing contracts, the reprocessors whose contracts would be delayed or canceled would probably also require compensation—perhaps by means of continued payments on the existing contracts (since those who were to receive plutonium would still be receiving plutonium without reprocessing). Delaying reprocessing of a decade's worth of spent fuel would require additional spent fuel storage either at reactor or reprocessing sites. All told, the subsidy required to financially compensate all the relevant parties in the reprocessing-deferral variant might be comparable to or higher than the subsidy required to burn plutonium in LWRs that would otherwise burn LEU, discussed above.31
Environment, Safety, and Health
In the variant involving a net expansion of MOX use, the ES&H impacts would be comparable to those described above in the case of U.S. LWRs. In the variant involving deferral of reprocessing, the net additional ES&H burden
would probably be smaller than that for other options, since the WPu would displace commercial plutonium that would be used in any case. As with other options, there would be some ES&H risks involved in the processing of the pits to oxide, and steps to minimize the risks of accidents during the international shipment would be required. But there might also be some ES&H benefits: workers at MOX fabrication facilities, for example, would be exposed to lower radiation doses from WPu than they would be from reactor-grade plutonium, and adding a decade or more to the time spent fuel would be stored prior to reprocessing would reduce the radioactivity of the fuel when it was eventually processed.
This “substitution" option sends a variety of signals, which vary depending on the specifics of the approach. Making use of potential excess MOX capacity would mean larger-scale use of plutonium fuels than would otherwise occur, which some would oppose. Critics of the use of separated plutonium fuels might see an approach that tied disposition of WPu to continued large-scale MOX operations as irrevocably confirming MOX plans that might otherwise be canceled, and as conferring the political legitimacy of disarmament on MOX operations. Parties interested in maintaining the momentum of commercial reprocessing might view the variant involving a deferral of such reprocessing until current plutonium surpluses are consumed as a fundamental threat to the economic viability of current reprocessing plants, and thus to the plutonium fuel cycle in general (see discussion in Chapter 6, "General Considerations").
The Spiking Option
The use of West European LWRs for the spiking option is also possible. This approach would appear to have little value, however, since there is ample capacity to burn the MOX fuel to full discharge burnups. Employing the spiking option in this approach would shift the burden of the resulting reduction in capacity factor to the West Europeans or Japanese, presumably requiring additional compensation.
The Elimination Option
West European or Japanese participation in the elimination option would require the additional utilization of West European spent fuel reprocessing facilities. Capital costs to construct commercial reprocessing facilities in the United States and the former Soviet Union would be avoided, although amortization of the capital cost of the commercial facilities would probably be included in the price charged for reprocessing services in other countries. British
and French reprocessing firms are actively seeking additional reprocessing contracts for the post-2000 period.
CURRENT-GENERATION LIQUID-METAL REACTORS
Description of Technology and Status
Current-generation liquid-metal reactors (LMRs) are sodium-cooled systems, fueled with HEU or plutonium fuels, typically with an enrichment of 2040 percent rather than the few percent used in LWRs. There are nine such reactors in the world. The development of the LMR was initiated by the U.S. Atomic Energy Commission in the late 1940s to exploit the potential of a fast reactor to generate more fissionable plutonium, by converting U-238 to plutonium, than it burned in producing power. It was thus called a "breeder reactor." Other countries, most notably France, Great Britain, Germany, Japan, and the Soviet Union undertook major programs in breeder-reactor development. The United States, Great Britain, and Germany have now canceled their breeder programs or scaled them back to a research phase. France is planning to convert the Superphenix, the world's largest LMR, to a research facility, focusing on plutonium burning rather than breeding, and is continuing its LMR research and development program. Japan has just started operating a new experimental LMR, Monju, but has decided to substantially postpone deployment of commercial-scale LMRs.
The current generation of LMR that has evolved from these development programs consists of a reactor core, a primary coolant system that circulates radioactive liquid sodium through the core to an intermediate heat exchanger, and a secondary coolant circuit that circulates nonradioactive sodium to a steam generator. Sodium coolant is used because of its excellent heat transfer properties and because it permits the reactor to operate in a fast-neutron spectrum—since it is not an effective moderator of neutrons. There are two types of cooling circuit configurations: (1) the loop type, in which the sodium coolant circuit is made of piping external to the reactor vessel; and (2) the pool type, in which the core and intermediate heat exchanger is placed in a large sodium-filled vessel (the "pool"), and the secondary sodium piped from the intermediate heat exchanger to an external steam generator.
The reactor core is typically made up of assemblies of oxide fuel rods— stainless steel tubes containing pellets of the oxide fuel. Some of the rods, called the driver fuel, contain either MOX pellets or medium-enriched uranium pellets, and others, the fertile fuel, contain natural uranium oxide pellets. Altering the location and relative loadings of these fuels can change the conversion ratio (the ratio of fissile material consumed to fissile material produced). Fuel burnup in the range of 100,000-150,000 MWd/MTHM is desirable to hold down repro-
cessing requirements. This burnup level has been demonstrated in test reactors, and is well above the requirements for WPu disposition to meet the spent fuel standard. The oxide fuel has proven reliable through extensive verification tests and substantial reactor operations.
Because of the need to avoid reactions between the sodium coolant and water or steam, the reactor and coolant circuits are enclosed in inerted compartments. The system operates at low pressure, making it easier to protect against loss of sodium cooling than a high-pressure system such as the LWR. Extensive testing has demonstrated that the system can be controlled effectively in a fast-neutron cycle, principally because of the negative Doppler coefficient of reactivity of the fuel.32 A variety of technical problems have been encountered in several countries, however, including sodium leaks and fires in some cases, and unexplained shifts in core reactivity in others. These have contributed to poor availability records for some LMRs, and to the worldwide delay in commercializing LMR systems.
Recycling of the plutonium has been intended for the LMR concept from the start, in order to utilize the additional fuel produced. Thus, the capability to accept MOX fuel assemblies is already a feature of the LMR concept. LMRs operating in the once-through mode instead of the recycle mode can, therefore, convert separated WPu into spent fuel. The mix of elements in the spent LMR fuel would be similar to that of spent LWR fuel, except that the fraction of plutonium remaining would be higher (because of the higher initial fissile loading), and the remaining plutonium would be closer to weapons-grade (because of the neutron-absorption properties of plutonium in a fast spectrum). The fuel could be disposed of in the same geologic repository to be used for LWR fuel (though a separate license for the fuel as an acceptable waste form for disposal would be required). The proliferation resistance of the end product would in most respects be similar to spent LWR fuel. The existing LMRs have predominantly operated in this once-through mode, so this would not be a new approach.
If desired, the LMR core mix of fissile and fertile material and reflector material can be arranged so that the ratio of fissile material created to fissile material consumed is much less than unity. This characteristic would allow the
LMR to be operated in the recycle mode as a net destroyer of the plutonium (in order to pursue the elimination option).
Reactor Capacity and Throughput: U.S. LMRs
LMRs have higher fissile loadings than do equal-capacity LWRs. The LMR capacity necessary to implement the spent fuel option on 50 tons of U.S. WPu over a 25-year period is little more than 1 gigawatt-electric (GWe).
The United States has two operable LMRs: the Experimental Breeder Reactor-II (EBR-II), located at the National Reactor Testing Station near Idaho Falls, Idaho, and the Fast Flux Test Facility (FFTF), located at the DOE complex in Hanford, Washington. EBR-II has a capacity of only 16.5 MWe and, therefore, could not make a substantial contribution to the disposition of the WPu. In early 1994, the Clinton administration decided to shut down the EBR-II as part of a broader decision to cancel U.S. support for advanced LMR research.
The FFTF does not have a steam generator, but instead transfers the heat from its secondary circuit to an air-dump heat exchanger. It was built to test LMR fuel and gain operating experience with a loop-type LMR system. The once-through mode that would be necessary for either the spent fuel option or the reactor-spiking option has been FFTF's operating mode to date. FFTF operated with very high reliability until the early 1990s, when reactor operations were stopped for budgetary reasons. A return to power would be practicable if a decision was made to do so, however.
FFTF has a capacity of 400 MWt, which also falls well short of the roughly 1 GWe needed for a U.S. disposition campaign. FFTF is big enough to perform the reactor-spiking option, however. A recent study by Westinghouse Hanford indicates that FFTF could spike 50 tons of WPu in less than 25 years. No significant reactor modifications are indicated to be necessary. The reactor-spiking option is only an interim solution, however, so a plan to use FFTF for reactor spiking would have to include the intended permanent solution, for which FFTF could make only a small contribution. In addition, FFTF produces no electricity, and hence would produce no revenue to offset the substantial costs of MOX fabrication and reactor operation.
FFTF is located on the same site as the incomplete FMEF MOX fabrication facility described in the first section of this chapter. FMEF could be adapted to fabricate assemblies at a rate that would support the above disposition of 50 tons of WPu in less than 25 years.
Reactor Capacity and Throughput: Foreign LMRs
There are six other operable LMRs in the world with capacities in excess of 100 MWe each. However, the three smallest have operated for roughly 20 years and, therefore, cannot be expected to make much contribution to the disposition
of the WPu. Interest therefore focuses on the three newest and largest, the French Superphenix, the Russian BN-600, and the Japanese Monju.
The French Superphenix is a 1,200-MWe plant, located at Bouvesse and operated intermittently since 1986. Its poor availability record was initially due to a leaking sodium storage tank and more recently to regulatory requirements to increase sodium fire protection and to evaluate a reactivity anomaly experienced at Superphenix's forerunner, the 233-MWe Phenix experimental LMR. These problems caused such an extended shutdown of Superphenix that it lost its license. After a prolonged relicensing process, the reactor was restarted in 1994. France now plans to use Superphenix primarily as a research facility to study actinide burning. If France's current plan were changed, Superphenix is just big enough, by itself, to implement the spent fuel option on 50 tons of U.S. WPu over a 25-year period, if it operated with much higher availability than it has in the past. France also has sufficient MOX fuel fabrication capacity to support use of Superphenix for this mission. Superphenix's past record gives little basis for confidence in future performance, however, and shipping WPu to this facility does not appear to have any major advantages over the more general substitution approach described above.
The Russian BN-600 is a 560-MWe plant, located near Yekaterinburg and operated since 1981. To date it has been fueled primarily with HEU, not MOX, although MOX fuel assemblies (including some WPu assemblies) have been tested in this reactor. MINATOM officials report that "a complete conversion of these reactors [the BN-600 and the older and smaller BN-350] to MOX fuel is not possible owing to their design and physical features" (Bibilashvili and Reshetnikov 1993, p. 32). If the BN-600 could be fully converted to MOX, its capacity is big enough for the reactor-spiking option, but it is only half that necessary to implement the spent fuel option on the nominal 50 tons of Russian WPu over a 25-year period. It is questionable whether the BN-600 could operate safely over that long a period. The BN-600's predecessor, the BN-350, was initially plagued by sodium leaks and fires; reports of a substantial sodium leak and resulting fire at the BN-600 in 1993 suggest that this problem has not been entirely resolved. A larger plant, BN-800, a 750-MWe pool-type, is contemplated, but construction of the two lead units has been halted for some years (see description in section "Advanced Liquid-Metal Reactors" below).
A new experimental 280-MWe LMR, Monju, went critical at Tsuruga, Japan, in the spring of 1994. Although Monju has its whole operable life ahead of it, the small capacity would allow this reactor to implement the spent fuel option on only 10-15 tons of U.S. or Soviet WPu over a 20-year period. Monju is big enough to implement the reactor-spiking option on either the U.S. or the Russian WPu, but all the issues described above for the more general case of substituting WPu for civilian plutonium would apply in this case as well.
In short, to carry out the spent fuel option, unless Superphenix achieved high availability, essentially the total world capacity of presently operable
LMRs would be needed. Yet many of these reactors are not presently operating and the largest plant, Superphenix, had a poor availability record when it was operating. With this experience, it would not seem to be prudent to count on these systems to provide the necessary capability. Transport of large quantities of WPu outside the United States and Russia would be required. The available capacity is even less adequate to implement the "elimination" option, for which LMRs with repeated recycle might be a candidate. In short, the panel does not recommend that the use of existing LMRs be pursued further as a major option for disposition of excess WPu.
CURRENT NAVAL AND RESEARCH REACTORS
All nuclear reactors on U.S. Navy ships use HEU fuel.33 Some Russian naval reactors also use HEU fuel. In principle, one could imagine that these HEU-fueled reactors might be modified to use plutonium fuel instead. Because somewhat more information is available, we focus on the U.S. case.
There are 158 operating U.S. naval reactors. All are low-power reactors: the average power is about 150 MWt. Naval reactor cores have long lives, with new cores planned to last the life of the ship or, at least, for decades. For example, the cores on the recently refueled aircraft carrier Enterprise are expected to last 20 years and those for the Nimitz are expected to last at least that long (Schmitt 1993b). The average use factor is only 10 percent of rated power, and naval reactor-core endurance (in full power days) and power density (in kilowatts/liter) are comparable to those of commercial reactors (Schmitt 1994). However, naval reactor cores withstand a much longer time of high temperature operation and orders of magnitude more large power transients than do commercial reactors. In addition, naval reactor cores must withstand more than 10 times the shock loading of commercial reactors.
The fuel fabrication facility for U.S. naval reactor fuel is being decommissioned because fuel is now available to support fabrication of cores through about 2001. A fabrication facility for naval fuel will not be needed until then, and the Navy expects to seek bids later this decade for such a facility. HEU is already available to cover projected requirements until about 2006. The Navy expects that beyond this time, HEU will come from U.S. weapons dismantlement (Schmitt 1993c). Given the uncertainty in out-year projections of the defense budget, future naval vessel construction and, therefore, the rate at which new reactor cores would be needed, are quite uncertain.
Since naval reactors do not now use plutonium-based fuels, any introduction of such fuels would require substantial development, not unlike that which has occurred over the last several decades in the case of plutonium fuels for commercial LWRs. Because of the extremely high reliability requirements for
naval fuel, the development time probably would be longer than in the case of developing a new commercial fuel type. The Office of Naval Reactors estimates it would take 20 years to qualify a new fuel system, develop new core designs, and develop the fuel manufacturing facility to handle plutonium (Schmitt 1993a). Although this may be an overestimate, because of the obvious reluctance of the Navy to plan to use plutonium in its cores, the required time is not likely to be substantially less.
According to the Office of Naval Reactors, the U-235 burnup per year is approximately 1.1 tons in the entire operating fleet and, therefore, about the same quantity of plutonium could be burned if the entire fleet were converted to plutonium fuel cores (a remote possibility). A larger amount of plutonium would be irradiated to the spent fuel standard each year, however. Since the Navy operation is based on extremely long-lived cores, any change to a shorter refueling cycle (so as to remove plutonium when it was sufficiently irradiated to be self-protecting, increasing the disposition rate) would require major restructuring of the Navy logistics and operations cycles.
In short, the use of naval reactors is not a promising option for plutonium disposition. One aspect of naval reactor experience that might be of use for U.S. plutonium consumption, however, is the technology that has been developed to produce very high reliability fuel. This technology may be of use if the United States chooses to develop MOX or plutonium cores for use in commercial reactors, or commercial-type reactors, for the consumption of the WPu. In particular, the naval reactor developments may be of use if the "elimination" option is selected. The technology is classified and, consequently, a high-level decision would be required for the technology to be usable in the commercial world or shared with Russia.
Current research reactors, like naval reactors, do not offer an attractive option for the disposition of WPu. Research reactors are generally small in capacity and in duty factor, they refuel only rarely or not at all, most were not designed for the use of plutonium fuel, and they are highly dispersed geographically and often located in institutional settings that would be difficult to safeguard. Given the availability of more practical possibilities, research reactors do not deserve serious consideration for the disposition mission.
ADVANCED LIGHT-WATER REACTORS
Description of Technology and Status
Advanced light-water reactor (ALWR) nuclear plants are being developed in the United States and other countries to meet future baseload electric-generation capacity needs. One or more ALWRs could be built for the plutonium disposition mission, although in general this option would appear to have higher initial capital costs and longer time-lines than the use of existing or partly
completed LWRs. A summary of these ALWR development programs, and the schedules set by the developers, is given below, followed by a discussion of the applicability of ALWRs to this WPu disposition mission.
Two main classes of follow-on ALWRs are under development. One approach is evolutionary, with designs conceptually similar to those of present LWRs, but with substantial improvements in safety, reliability, and cost, derived from experience with the current generation of plants. These plant designs tend to be in the 1,300- to 1,500-MWe power range. Another approach involves a greater technological reach, in order to emphasize the concept of passive safety. These plants tend to be smaller, in the 600-MWe range.
U.S. organizations are developing three evolutionary designs with power output of 1300 MWe: an advanced boiling-water reactor (ABWR), being developed by General Electric (GE) (Redding and McGregor 1993); a pressurized-water reactor (PWR) known as the System-80+, being developed by ABB-Combustion Engineering (Turk and Matzie 1992); and another advanced pressurized-water reactor (APWR) being developed by Westinghouse, with designs of both 1,050 MWe and 1,300 MWe (Nuclear News 1992b). In the same power class, the French are developing a 1,450-MWe PWR, called N-4 (NEI 1985), and are collaborating with the Germans (Framatome and Siemens) to develop another 1,450-MWe PWR, called EPR (Baumgartl and Watteau 1994), for export. Great Britain is constructing a 1,300-MWe PWR, called Sizewell B (NEI 1988). The Japanese are constructing two 1,300-MWe GE-ABWRs.
In these evolutionary plants, safety is improved by: (1) reducing the probability of core-damage accidents by incorporating additional passive features and modern control-room technology that reduce the burden on the operator in handling abnormal conditions; and (2) improving severe accident mitigation through more rugged decay-heat removal systems and containment, and passive containment temperature and pressure control. Reliability is increased by utilizing improved component design and materials derived from operating experience. Costs are projected to be reduced by simplified configurations and shorter construction time through automated, integrated management systems and modern modular construction processes.
The fuel and fuel reloading systems are essentially identical to those in the present plants, since experience with LWR fuel has been quite favorable. The ABWR and the System-80+ were designed to be able to utilize a full-MOX core. Most of the other designs could handle a one-third MOX core without significant modifications, might be able to utilize a full MOX core without such modifications, and certainly could be modified for full-MOX if that proved necessary. Other evolutionary designs could be modified to do so (GE 1994, Westinghouse 1994).
Two ABWR plants have been authorized in Japan and are in the early stage of construction, with completion scheduled for 1997-1998. A forerunner of the System-80+—a System-80 with a number of the 80+ features incorporated—is
under construction in Korea, scheduled for completion in 1996. The System80+, the Framatome, and the Sizewell B machines are being offered in competition for construction of a reactor in Taiwan. Both the ABWR and the System 80+ designs have been submitted to the NRC. Final design approval was given in 1994, and design certification is expected in 1996. The design of the APWR is being finalized in Japan. Three N-4 plants are under construction in France and are scheduled for completion by 1995-1997. The Sizewell B plant is under construction in Great Britain and is scheduled for completion in 1994.
Designs of smaller-size plants are also under development. In the United States, three 600-MWe designs are underway: AP-600 (Westinghouse) (Bruschi and Andersen 1991), the Simplified Boiling-Water Reactor (SBWR) (GE) (Redding and McGregor 1993), and the Safe Integral Reactor (SIR) (ABB-CE 1989). These systems utilize passive emergency cooling features instead of active, high-powered equipment for that purpose. Their designers argue that this will lead to improved safety, greater simplification and economy, and reduced demand for emergency response by the operators. Improved safety, reliability, and cost are also achieved through changes similar to those for the larger plants outlined above. Although the power-generating portions of the designs are conceptually the same as in the current generation of LWRs, the emergency cooling features need extensive testing and represent a new technical element in the licensing process. Both the AP-600 and the SBWR designs have been submitted to NRC for certification, and final design approvals are expected in 1996 and 1997—three years behind the evolutionary plants.
The Japanese are developing a larger unit, nominally 900-MWe output, called the Simplified Pressurized-Water Reactor (SPWR) (Sako et al. 1992), with passive safety features patterned after the AP-600. The design is at the conceptual level, and plans have not yet been formulated for detailed design and licensing. A 600-MWe ALWR with passive safety features, the VPBER-600 (Mitenkov et al. 1992), is at an early stage of development in Russia. There is no schedule for authorization or construction.
A passive ALWR design is also being developed in Sweden (ABB-Atom), called PIUS (Process Inherent Ultimate Safety) (Nuclear News 1992b). NRC has completed a limited portion of a pre-application review of PIUS, deferring further review until an application for design certification is received. PIUS utilizes passive safety features to a much greater degree, involving the power generation functions as well as the emergency cooling functions. This system is at an early stage of development and has limited funding, so that the time horizon for deployment is estimated to be 2010 and later. Systems similar to PIUS, at an even earlier conceptual phase, are being developed in Japan.
In general, total electricity costs for these smaller passive systems are estimated to be somewhat higher than for the large evolutionary ALWRs, and there is greater uncertainty in the cost estimate and schedule because of the newer
features in these plant designs. The licensing of the plants will probably be more time-consuming because of the newer plant design features.
Reactor Throughput, Once-Through Fuel Cycle
The reactor throughputs for ALWRs would be similar to those described above for the case of LWRs of existing design. New systems built for plutonium disposition, however, could be built from the outset with the features needed to accommodate full-MOX cores with high plutonium loadings. One of the larger class of ALWRs, using fuel with an enrichment in the neighborhood of 6-7 percent plutonium (with burble poisons), operating with a 75-percent capacity factor and an average burnup of 42,000 MWd/MTHM, could process all of the nominal 50 tons of excess WPu in 30 years. Two of the smaller class of LWRs, amounting to a comparable total energy output, would be required to process a comparable amount of plutonium over a comparable period.
ALWRs, being in the yet-to-be-built category, could not be made available to receive fuel containing WPu as early as could existing reactors. Even if high national priority were given to building them, they would not become available in less than 10 years. The delay in potential initial fuel-loading date in that case would be less than 10 years, however, because even existing reactors would need an adequately sized MOX fuel-assembly fabrication capability to be made available, and a lengthy approval and licensing process would be involved.
The alternatives for fuel fabrication for ALWRs are the same as those described above for LWRs of existing design.
Approvals and Licenses
Because the ALWR designs have not yet been approved by the NRC, there would be some additional licensing uncertainty compared to the use of LWRs of existing designs. The safety and environmental impact of the ALWRs, however, has received major emphasis in their design, so that they should be measurably safer than the current-generation U.S. LWRs. The major regulatory issues in the United States would be the passive safety features of the smaller plants and the utilization of these designs for processing WPu. Construction of any new nuclear-power plants in the United States, however, would be likely to raise public controversies. For the spent fuel option, reprocessing would not be required, and the existing French commercial implementation of plutonium recycle and the limited earlier demonstration in the United States could provide an experience base for this mission in ALWRs.
The above considerations apply to all ALWRs, but further evaluation requires distinguishing the different ALWR design concepts. The large ALWRs
can be classified as mature designs, with some models already under construction overseas. They have met licensing requirements (with LEU fueling) in Japan and Great Britain, and they are nearing the end of the arduous process of NRC certification with LEU loading. The more advanced passive designs, by contrast, have considerably further to go in the licensing process, and therefore face greater licensing uncertainty. NRC or Defense Nuclear Facilities Safety Board approval of the MOX fuel fabrication facility will also be a challenge.
Earlier regulatory reviews of reactor-grade plutonium recycle in LWRs, such as the Generic Environmental Statement on Mixed Oxide fuel (GESMO), provide a significant technical foundation for the licensing process, but that work was halted more than 15 years ago and focused on the one-third MOX loading option. The regulatory reviews required for the ALWR option would probably focus on full-MOX loadings and the impact of changes in safety and environmental rules since the GESMO reviews ended. The use of erbium oxide as a burble poison to increase plutonium loading would add another new feature in licensing.
Despite these uncertainties, one possible advantage of construction of a new ALWR for this mission over use of an existing LWR is that the new system could be built on an existing government site (presumably co-located with other steps in the disposition process, such as fuel fabrication and pit processing). This could reduce problems of public acceptance, transport, and handling of plutonium at multiple sites, and the like.34 Whether the overall approval process would be more uncertain and longer for ALWRs or for existing LWRs depends in significant part on the specific sites, reactor designs, and politics involved. In either case, proposals for the use of plutonium in U.S. reactors can be expected to engender a significant debate.
Safeguards, Security, and Recoverability
The issues here would be the same as those described for the case of LWRs of existing design. As just noted, ALWRs built for this mission would presumably be based at a single government site, minimizing transportation risks.
The initial capital cost of building a new reactor would inevitably be higher than the cost of utilizing existing reactors or finishing partly completed facilities. Some private vendors, however, are proposing concepts in which the initial capital cost would be covered by the private sector, in return for a guarantee of an annual plutonium disposition fee from the government sufficient to make the
electricity from the reactors competitive with other sources of electricity in the area where the reactor was deployed. In general, it appears that the overall net discounted present cost of new reactor approaches would be somewhat higher than existing reactor approaches (because of the higher cost of nuclear power compared to other generating options in the projected U.S. market over the next 40 years), but there is considerable uncertainty in these calculations. (For a detailed economic comparison, see Chapter 6.)
Environment, Safety, and Health
As noted above, the ALWRs are designed to be measurably safer than current-generation LWRs. Current LWRs in the United States, however, are believed to be acceptably safe, and as noted above, if appropriate modifications were implemented, use of full-MOX cores should not reduce existing safety margins. (For a detailed discussion, see Chapter 6.)
The perceived signals relating to fuel-cycle policy in this case would likely be roughly the same as for LWRs of existing design.
The Spiking Option
The spiking option puts high priority on early denaturing of WPu, and therefore does not fit well with alternatives such as the ALWR, which require building new reactors for that purpose. The utilization of existing LWRs for spiking could start the disposition campaign sooner, with a campaign period no longer than one using ALWRs.
The Elimination Option
The characteristics for this option in ALWRs would not differ from those described above for current LWRs.
ADVANCED LIQUID-METAL REACTORS
Description of Technology and Status
Three alternative designs of advanced liquid-metal reactors (ALMRs) are under development outside the former Soviet Union, and another has been developed there. All of these reactors are inherently designed to utilize plutonium fuels and thus are obvious candidates for plutonium disposition. Three utilize the current MOX fuel, and one is based on a metal alloy of uranium, plutonium, and zirconium. The three oxide fuel versions are being developed in Western Europe, Japan, and Russia, and the metal fuel version in the United States. Re-
processing and recycle of plutonium is an inherent part of the operating concept of these reactors. Such reprocessing and recycle is applicable to the elimination option, but not to the spent fuel option. If operated in a once-through mode, however, ALMRs could be used to transform WPu into spent fuel, just as other reactors could. The capital costs of these liquid-metal reactor concepts are expected to be higher than those of LWRs, however, and they are much farther from being licensed in the United States than are evolutionary ALWRs. Hence these reactors are of greater interest for the long-term elimination option than for the spent fuel option.
The U.S. design is being developed by General Electric, under the sponsorship of the U.S. Department of Energy, and is based on the metal-alloy fuel pin design being developed by Argonne National Laboratory in Illinois.35 In 1994, government support for development of this reactor concept was canceled, but it could be restarted if this concept were chosen as a plutonium disposition option or a future electricity source. The overall concept, called the Integral Fast Reactor (IFR) (Till and Chang 1988), is a significant advancement over current LMR designs. The plant design is based on the pool configuration described above in the section “Current-Generation Liquid-Metal Reactors."
The use of metal fuel results in lower reactor outlet temperature (500° C), steam temperature (430° C), and net plant efficiency (36 percent) compared to the other two ALMR alternatives, but it is projected that this loss in efficiency will be balanced by superior fuel-cycle economics. A smaller reactor power output, about 303 MWe, permits the incorporation of a passive decay-heat removal system. Each reactor drives a single steam generator, and two reactor/steam generator units would be coupled through their steam loops to drive an economically sized, 606-MWe (net), turbine.
The most significant advance in the U.S. IFR program is a pyroprocessing approach intended to significantly reduce the costs, wastes, and proliferation risks of reprocessing (compared to the conventional aqueous process used with oxide fuel).36 In this integrated reprocessing approach, the plutonium is never fully separated in a form that could be used directly in nuclear weapons, thereby reducing safeguards concerns.37 The metal fuel lends itself to this pyroprocess
ing approach. To further increase proliferation resistance, the power plant, reprocessing facility, and MOX fabrication facility are co-located.
The design employs modular factory fabrication to minimize capital cost in an effort to offset the loss of economy of scale. Other distinctive features include seismic isolators supporting the nuclear steam supply system and electromagnetic pumps driving the primary sodium. A comprehensive test program has been defined for demonstrating the newest features of the plant design. A test program has also been underway for several years at the Argonne National Laboratory to demonstrate the performance characteristics of the metal fuel pin and of the pyroprocess that would be used for processing the spent LMR fuel in the recycle mode that ALMRs would operate in commercial deployments. Licensability reviews have been completed favorably with the NRC. NRC design certification has not yet started; prior to program cancellation, design certification was scheduled to be completed by 2005. Construction of a full-size prototype reactor co-located with a reprocessing and fabrication facility was planned if the demonstration tests were successfully concluded and after completion of preliminary and final design.
The West European design, known as the European Fast Reactor (EFR) (Ebbinghaus et al. 1992), is the result of a collaborative effort, mainly between France, Germany, and Great Britain. Here, too, government decisions have put the program's future in doubt, as both Britain and Germany have withdrawn their support. The design is for a large, 1,450-MWe plant with the pool configuration, of the same concept as the current generation of oxide-fueled LMRs, incorporating substantial improvements in safety, reliability, and economy derived from construction and operation of Western Europe's existing LMRs. The newest features of the design have been the subject of a comprehensive test program. The oxide fuel assemblies are to have an in-core residence time of six years and to achieve a peak burnup of 200,000 MWd/MTHM. The core is I meter high and 4 meters in diameter. The large-capacity primary sodium circuit drives six parallel secondary sodium circuits each containing a steam generator. A reactor outlet temperature of 545° C permits a steam temperature of 490° C and a resulting net plant efficiency of 40 percent.
The Japanese design, known as the Demonstration Fast-Breeder Reactor (DFBR) (Miura et al. 1992) is for a medium-size, 670-MWe plant. The Japanese government has recently indicated that construction of such a commercial-scale breeder will now be postponed for a decade or decades. The DFBR is a sodium loop design. The loop components are interconnected by top entry piping that projects down into free sodium surfaces in the components, thereby precluding a siphoning risk that would otherwise exist with the loop configuration. This medium-capacity primary sodium circuit drives three parallel secondary sodium circuits, each containing a steam generator. The reactor outlet temperature, steam temperature, net plant efficiency, and fuel burnup are slightly higher than those described above for the EFR. The details of the DFBR design are substan-
tially influenced by the high seismicity of Japan. The newest features of this design also have been the subject of a comprehensive test program. In addition, some Japanese officials have suggested that an international group fund a special-purpose LMR to be built in Russia to consume excess WPu. The issues associated with this concept are similar to those of other ALMR concepts, described in this section.
The Soviet Union also designed a follow-on LMR, known as the BN-800 (Minkov et al. 1990). MINATOM hopes to build four of these 750-MWe plants, three at the Mayak complex at Chelyabinsk. Construction on two of these plants started in the 1980s, but has been stopped for several years because of lack of funds. The BN-800 is an important special case of follow-on LMRs, because use of plutonium fuel in these reactors is MINATOM's preferred option for disposition of both excess WPu and civilian separated plutonium (now building up in storage as a result of continued reprocessing at the Mayak site) (Mikhailov et al. 1994). It is therefore worth describing this design in some detail.
The BN-800 design is similar in many respects to the BN-600, described above. The major BN-800 reactor components are of the same size and comparable in design to those of the BN-600 reactor, with the major differences being in the secondary systems. The most significant difference is the intended fuel. The BN-800 reactor is designed to use a full core of MOX fuel, while the BN-600 uses primarily uranium oxide, and, as mentioned above, is described by MINATOM as being incapable of shifting to full-MOX fueling. Raising the output of the reactor design from 560 to 750 MWe required an increase of the core volume; this was accomplished by increasing the core height from 0.75 to 0.95 meters, and by increasing the number of fuel assemblies, which increased the core diameter from 2.05 meters to approximately 2.25 meters. Despite the increase in core diameter, the radial dimensions of the central section of the reactor were kept unchanged, thanks to a reduction of the thickness of the radiation shield and a decrease in the thickness of the radial blanket. The BN-800 core is designed to have three radial zones of differing enrichment, one more than in the BN-600 core.
Some important changes were made in the reactor safety provisions, resulting in part from new guidelines on nuclear safety adopted by the Soviet Union in 1982. The most important changes were:
Increasing the number of control rods from 27 to 30, making it possible to create two independent and redundant reactivity control systems.
Introducing technical means (based on acoustic and neutron noise) to detect the onset of sodium boiling (which could lead to overheating because of the positive sodium void coefficient).
Providing operational guidance and procedures to the operator on what to do to prevent accident development.
Placing the entire equipment of the primary and secondary circuits on a monolithic concrete slab to improve seismic safety. This includes the steam generators, the refueling equipment, the auxiliary systems for operating the components with liquid metal, the heat transport equipment for the primary and secondary circuits, and the safety subsystems.
Adding a new emergency cooling system involving auxiliary sodium-air heat exchangers connected in parallel with the steam generators in each secondary circuit. This system provides for emergency residual heat removal during accidents initiated by the loss of power supplied to the reactor subsystems or the loss of feedwater (for example, in the case of a rupture in the water-steam equipment or piping). In all other accidents, the emergency cooling of the reactor is carried on through the steam generators.
Adding a redundant standby control room allowing the reactor to be shut down, and its main neutronic, thermal, and fire-safety conditions to be monitored, in the event the main control room becomes unusable.
Setting up the reactor scram-rod system as two independent subsystems, each being able to actuate independently the insertion of all the rods by the force of gravity. The independence of the two scram-rod subsystems is ensured by placing their components in different rooms, by laying communication lines along different cable passages, and by connecting each of the two subsystems to a different power supply source. Each subsystem operates on three independent information channels.
Thus the safety of the BN-800 design is significantly better than that of the BN-600 that is now operational. MINATOM officials report that the design has received licensing approval. Whether the design incorporates adequate redundancy in its safety features to meet international safety standards, however, requires additional information and further analysis.
In addition to this sodium-cooled design, Russian institutes are carrying out conceptual design studies of ALMRs cooled by molten lead, or a molten lead-bismuth mixture, rather than molten sodium. Conceptual designs have been developed for both a 300-MWe modular unit patterned after the PRISM-IFR (Gleukler and Quinn 1994) and a 1,000-MWe unit patterned after the evolutionary ALMRs. A significant advantage of the lead-cooled reactor is the elimination of the risk of sodium leaks and fires, since lead is inert to both air and water. This allows the secondary coolant loop and the intermediate heat exchanger to be eliminated, simplifying the system. Lead has heat transfer properties that are superior to those of sodium, including a larger margin between the operating temperature and coolant boiling temperature, increased natural convection potential, and a larger system thermal inertia. The proposed fuel is a high-density mixture of uranium and plutonium mononitrides (PuN-UN) with a ferritic steel
cladding. The gap between the cladding and fuel is filled with molten lead to reduce fuel-clad interaction and increase heat transfer.
This lead-cooled ALMR is still in the conceptual design phase. Corrosion and structural properties of the plant material will have to be defined and adequate reliability demonstrated. The proposed nitride fuel form will require extensive testing and qualification. It is highly unlikely that this concept could be developed and deployed in less than two to three decades.
Reactor Throughput, Once-Through Fuel Cycle
As described in the above section "Current-Generation Liquid-Metal Reactors," the ALMR capacity necessary to implement the once-through spent fuel option would be about 2 GWe to process a combined U.S./former Soviet Union total of 100 tons of WPu over 25 years. The precise capacity necessary would depend mainly on the reactor design and on the degree of fuel burnup necessary to achieve the desired level of proliferation resistance.
ALMRs, being in the yet-to-be-built category, could not be made available to process WPu as early as could existing reactors, and the costs and uncertainties involved in their use would be higher. Capital costs in the billions of dollars would be needed to provide the capacity to process 100 tons of WPu in the spent fuel option. For ALMR designs such as the IFR, a prototype plant would probably have to be built first. In the United States, prospects for gaining approval for construction of a new LMR must be judged to be highly uncertain. Even once the requisite licensing and political approvals had been gained, a new LMR would take a minimum of 10 years to complete in a "crash" program. An ALMR design such as the IFR would take longer.
The time required to complete the BN-800 reactors in Russia is even more difficult to predict. The fundamental barrier to completion of these reactors is lack of available funds; safety reviews and the need to gain political and licensing approval could also delay completion. Particularly given the substantial supplies of low-cost uranium and enrichment services available in Russia, the electricity cost from reactors of this type is likely to be significantly higher than the cost of electricity produced by LEU-fueled LWRs, or the cost of other sources of electricity; in the current economic environment in Russia, such a large-scale subsidy may be difficult for MINATOM to justify. Completion of these reactors in the near term appears unlikely in the absence of substantial Western assistance. Because of the decline in industrial production (which will probably not recover completely for a substantial period), moreover, electricity demand has declined, and it is by no means clear that new electrical capacity is needed in the areas planned for these reactors. Because of factors such as these, some top MINATOM officials have acknowledged that the first BN-800 is un-
likely to be operational for at least 10-15 years.38 Even this estimate appears optimistic; it is difficult to rely on the availability of these facilities on any set schedule. Thus the BN-800 does not appear to meet the important criterion of minimizing the total time before disposition of plutonium could be accomplished.
We note that in saying it may not be desirable to wait for the BN-800s for plutonium disposition, we are neither endorsing nor rejecting the BN-800s as an option for Russia's future energy supply. The 30 tons of civilian separated plutonium already in stock is more than sufficient to operate the four BN-800s, regardless of what is done with the WPu: each reactor only requires 2.3 tons of plutonium as startup fuel (meaning that less than 10 tons is required to start all four that might be built), and each produces more plutonium than it consumes thereafter.39 To use both the 30 tons of civilian separated plutonium and the nominal 50 tons of excess military plutonium in these reactors would mean continuing to add more fresh plutonium as spent fuel is removed, rather than allowing the reactors to fuel themselves through reprocessing and recycle of the plutonium they produce, as they were designed to do. The net result would simply be a much larger quantity of spent fuel awaiting reprocessing in the fuel cycle for these reactors.
Given the long schedules for providing ALMR reactor capacity, the availability of fuel fabrication capacity would probably not be a limiting factor governing the schedule for initial fuel loading in the ALMR option. As noted above, the United States already has an incomplete LMR MOX fuel assembly fabrication facility, the FMEF at Hanford, which could be completed, modified for current safeguards and environmental standards, and used to produce fuel at a rate that would support the disposition of 50 tons of U.S. WPu in less than 25 years in the spent fuel option. A facility comparable to the U.S. facility could be provided in Russia by completing the Chelyabinsk MOX fabrication plant, which was designed to provide fuel for the BN-800s.
Approvals and Licenses
As noted earlier, gaining political and licensing approval for construction and operation of an ALMR in the United States would be difficult, because of the domestic politics surrounding reprocessing and fast reactors. The current government policy is to terminate the fast-reactor program. This option therefore has substantial licensing and approval uncertainties. Indeed, consideration of the institutional issues and uncertainties involved suggests strongly that the U.S. government is the only viable organization that could sponsor and fund an ALMR deployment in the United States.
In Russia, the BN-800 design is reported to have received licensing approval. The national and local politics that would surround a decision to complete the BN-800s are difficult to predict, given the fast pace of political changes in that country. It seems likely, however, that if money was available, gaining the needed approvals would be less difficult in Russia than in the United States.
Safeguards, Security, and Recoverability
The safeguards and security issues in the use of ALMRs on a once-through cycle for the spent fuel option are not fundamentally different from those for other reactor types. Because of the high fissile loading in the fuel, a smaller total amount of fuel would have to be fabricated, but each unit of that fuel would contain more plutonium. In the case of the IFR, there are options for incorporating the WPu with various minor actinides and other radioactive species in the course of fuel fabrication, which could reduce the risk of theft of the material during the process, though it may at the same time complicate accountancy and require a somewhat modified safeguarding approach (Wymer et al. 1992).
As noted in the discussion of current LMRs, the mix of elements in the spent LMR fuel would be similar to that of spent LWR fuel, except that the fraction of plutonium remaining would be higher (because of the higher initial fissile loading), and the remaining plutonium would be closer to weapons-grade (because of the nonfission capture properties of plutonium in a fast spectrum). The proliferation resistance of the end product would in most respects be similar to spent LWR fuel.
In MINATOM's concept for plutonium disposition, the plutonium in the BN-800 spent fuel would ultimately be reprocessed and reused. Thus, while the WPu would initially be embedded in highly radioactive spent fuel (as in other spent fuel options), it would then be separated again. Only a few tons would exist in separated form at any one time, however. The BN-800 as currently conceived does not incorporate the integral reprocessing approach envisioned for future U.S. LMRs and thus raises greater safeguards and security concerns.
The capital costs of ALMRs are expected to be significantly higher than the costs of comparable LWRs. Since capital charges are a large fraction of the total cost of nuclear-generated electricity, the overall electricity cost from such reactors is also likely to be higher.
Plutonium would be a less costly fuel for ALMRs than uranium (in contrast to the LWR case), because of the higher costs for uranium purchases and enrichment for their more enriched fuels. But that would not make the total subsidy required for plutonium disposition (compared to the production of electricity by other means) less than it would be in the case of LWRs.
Environment, Safety, and Health
Few of the ES&H aspects of ALMRs are fundamentally different from those of other reactor types. The ALMRs outside the former Soviet Union are designed to offer high reactor safety and significantly lower operational exposures than LWRs. Confirming the safety of the BN-800 design would require more information than is available to the panel and further analyses.
If ALMRs were operated on a once-through fuel cycle, the spent fuel would ultimately have to be disposed. LMR spent fuel has not yet been certified as an acceptable waste product for disposal—in part because these systems and their fuels were largely designed with reprocessing rather than direct disposal in mind. The higher fissile content of LMR spent fuel would lead to greater long-term criticality concerns for repository disposal than in the case of LWR spent fuel. It is doubtful whether the metal fuel used in the IFR would be an acceptable waste form for the oxidizing environment in the proposed U.S. Yucca Mountain repository. A waste package could in principle be designed to attempt to protect the fuel from the oxidizing environment, but to date the U.S. waste disposal program has avoided placing primary reliance on such engineered barriers.
The U.S. decision to cancel government support for the IFR program was based in part on a desire to send a signal that the United States did not support a near-term transition to the plutonium fuel cycle for which fast reactors were designed. Construction of a fast reactor for plutonium disposition would be interpreted by some as sending the opposite signal, which would be contrary to current U.S. policy. Assistance for completion of the BN-800 reactors in Russia would send a similar signal and would provide a boost to the plutonium economy there. If operated for plutonium disposition on a once-through cycle, however, ALMRs would not inherently need to raise the issue of reprocessing and recycle. (See discussion in Chapter 6, "General Considerations.")
The Spiking Option
The spiking option puts high priority on early denaturing of WPu and therefore does not fit well with alternatives such as the ALMR, which require building new reactors for that purpose. The panel would not recommend using ALMRs for the spiking option.
The Elimination Option
LMRs, with their fast-neutron spectrum, can fission all isotopes of plutonium and are frequently put forward as a prime candidate for nearly complete plutonium elimination. Several countries are examining their potential as actinide burners. As noted above, some ALMRs, such as that being researched in the United States, employ an integral reprocessing technique in which the plutonium is never fully separated, mitigating some of the safeguards concerns that would otherwise arise from the repeated reprocessing and recycling required for the elimination option.
The fast-neutron spectrum of ALMRs requires a large fissile loading to achieve criticality, given the smaller fission cross-sections in a fast spectrum. As noted in the above section "U.S. Plutonium in Current-Generation U.S. Light-Water Reactors," after the initial period of reactor burning in an elimination option, the stock of plutonium would eventually be reduced to the point where it would no longer be sufficient to maintain criticality in the reactor. The remaining inventory would then have to be burned down exponentially by adding additional fissile material to maintain criticality in the reactor. For an ALMR, the in-core inventory that would have to be burned down in this manner is quite large, so longer times would be required to achieve very high destruction fractions than would be the case in a system with a smaller in-core inventory.40
As with other elimination concepts, pursuing this option would require an additional large capital investment to provide reprocessing capability. Although the pyroprocessing approach proposed for the IFR might prove to require lower capital costs than aqueous reprocessing, the uncertainties of cost estimates are particularly large since all previous production experience has been with aqueous homogeneous reprocessing. There are commensurate uncertainties in when such a capability could be provided.
Many of the other concerns regarding such elimination options described in the section on U.S. light-water reactors would apply in this case as well. In the case of ALMRs using an integral reprocessing approach in which the plutonium was never fully separated, such as was proposed for the U.S. IFR program, safeguards and security concerns would be substantially reduced compared to an elimination campaign based on the PUREX process.
MODULAR HIGH-TEMPERATURE GAS-COOLED REACTORS
Description of Technology and Status
The high-temperature gas-cooled reactor (HTGR) is a unique thermal-reactor approach. In the HTGR (or MHTGR, as some recent modular designs are called), the fuel would be encased in tiny particles or pellets, which would provide the first line of containment against release of fission products. This fuel pellet design offers the possibility of higher burnup on a once-through cycle than other reactors. These pellets would be within a graphite structure that would serve as moderator. The whole system would be cooled by helium. The hot helium could be used to produce steam, or, in the most recent designs, to drive turbines itself, in a so-called "direct-cycle" approach. In July 1993, this less mature direct-cycle system, known as the gas-turbine modular helium reactor (GT-MHR), was chosen as the basis for further commercial development in the United States, and it is the system now proposed by the developer, General Atomics, for plutonium disposition as well.41
In the General Atomics design, the pellets would be bonded together to form fuel rods, called compacts. The compacts would be inserted into vertical holes in hexagonal graphite blocks, called elements. (In the German and Russian HTGR designs, the graphite blocks are in the form of tennis-ball-sized spheres, and the reactor can be refueled online without powering down.) The General Atomics MHTGR core is an assembly of these blocks arranged in an annular shape. The center and outer portions of the core would be made from unfueled reflector blocks. Within the core, center reflector blocks are surrounded by 102 fuel blocks, which in turn are surrounded by outer reflector blocks. Each block can be removed by a fuel-handling machine (a feature that becomes important in one of the proposed plutonium-burning modes). The helium coolant flows downward through holes in the fuel blocks to a plenum at the bottom of the core. Metal-clad, borated graphite control rods control core reactivity. The reactor core is housed in an uninsulated steel reactor vessel, approximately 24 feet in diameter and 73 feet high, which is located below ground (GA 1994).
The kernel contains enriched uranium in the commercial design. For the plutonium disposition program, it would contain a mixture of plutonium oxides corresponding to 1.61 oxygen atoms for each plutonium atom. The kernel is surrounded by successive layers of a porous carbon buffer, an inner isotropic pyrolytic carbon layer, a silicon carbide barrier coating, and an outer isotropic pyrolytic carbon layer. These four layers together are called a TRISO coating. The TRISO plutonium particles are estimated to be 0.675 mm in diameter. This
is smaller than the particle previously planned for the version of the HTGR design developed for the New Production Reactor competition, a U.S. Department of Energy program (now canceled) to develop a reactor to produce tritium for weapons production. This particle had three additional layers, but experienced failures in testing.
The fuel designed for commercial application consisted of two types of kernels: fuel particles with 19.9-percent enriched uranium oxycarbide and fertile particles of thorium oxide (Nuclear News 1992a, p. 78). The plutonium-burning fuel would replace the uranium with weapons-grade plutonium. A high plutonium loading (88 percent) was used for fuel particles irradiated in the Peach Bottom MHTGR in the late 1960s and early 1970s. Performance was similar to that for HEU fuel, although sensitive to the initial oxygen-to-plutonium ratio (GA 1992, p. 4).
The MHTGR proposed by General Atomics in Phase I of DOE's study of reactors for plutonium disposition is based on gas-cooled designs that have been under development and, in some cases, in operation over the last 30 years. Two commercial plants have operated in this country: a helium-cooled 40-MWe demonstration unit at Peach Bottom, Pennsylvania, which operated from 1967-1974, and a 330-MWe plant at Fort St. Vrain, Colorado, built for commercial generation of electricity. Consistently poor performance led to the closing of that plant in 1991. A thorium high-temperature prototype reactor was run in Germany, but closed in 1989 due to technical problems.
The MHTGR design was developed initially for application as a commercial reactor to produce electricity. The plutonium burner was based on a reference design that was being developed by General Atomics as one of the two competing designs for the New Production Reactor. Both the commercial MHTGR and the plutonium-burner version used steam generators. General Atomics now has proposed that a direct Brayton cycle be used for both versions. The commercial reactor is labeled GT-MHR and the plutonium disposition reactor proposed in Phase II of DOE's study is labeled PC-MHR (plutonium-consumption modular helium reactor). This design also is the basis for a 1993 joint General Atomics/MINATOM agreement to work together to develop an HTGR for plutonium disposition, with the hope of funding for the project from the U.S. government.
The primary differences between the MHTGR proposed in Phase I of DOE's study and the PC-MHR proposed in Phase II are that the PC-MHR utilizes (Blue 1993):
direct cycle, using the heated helium in the reactor to feed directly into a gas turbine, with inlet temperature of 849º C;
magnetic bearings for the turbines;
compact plate heat exchangers, leading to a significant reduction in volume and an increase in efficiency; and
With these features, General Atomics estimates that this design could achieve a very high thermal efficiency, in the neighborhood of 46-47.7 percent (Schleicher et al. 1992; GA 1994, p. 2-52).
A significant difference between the GT-MHR and the PC-MHR proposed for plutonium disposition is that the commercial version has a vented, low-pressure containment, while the PC-MHR has a high-pressure, low-leakage containment. The MHTGR that had been under review by the NRC also differed from the MHR concept in that the MHTGR was air-cooled and the MHR is water-cooled.
General Atomics concludes that a plutonium TRISO fuel particle development and fabrication program would add a year and a half to the schedule that had existed for the New Production Reactor program. Since particle behavior for the large number of particles required for HTGR cores has been the critical safety issue (National Research Council 1992), however, significantly longer than 18 months might be necessary for successful demonstration of the plutonium particle fabrication process. Since the fuel remains the same for the PCMHR as for the Phase I concept, General Atomics' previous discussion of the fuel remains relevant. General Atomics notes that: "The critical path activity on the schedule for deployment of the PC-MHR is qualification testing and fabrication of coated plutonium fuel particles" (GA 1994, p. 13-2).
According to General Atomics, the main elements for the plutonium fuel development program are (GA 1992, pp. 19-20):
Design of a coated plutonium fuel particle.
Demonstration of plutonium fuel particle and compact fabrication capability, including process development, equipment design, and production scaleup.
Model development to predict plutonium fuel performance during normal and off-normal reactor service conditions.
Plutonium fuel irradiation performance and fission product behavior testing to validate fuel design, qualify fuel fabrication capability, and provide data for validation of fuel performance/fission product transport design methods.
In its review of the General Atomics proposal, DOE noted that six of the ten areas identified by General Atomics as needing development are related to the use of plutonium fuel. DOE concluded that the plutonium-burning HTGR technology "is not yet mature, and considerable development and testing remains to be completed" (USDOE 1993, Vol. II, p. SC 4&5-21). The Phase II report pre-
pared by General Atomics lists 22 plutonium fuel-design development needs. This report also identifies 28 facilities in the United States, Russia, France, and Japan that could in principle be used to support the design of the PC-MHR, and 7 new development and test facilities that would be required (GA 1994, pp. 3-48 to 3-57, 3-59 to 3-67).
The Lawrence Livermore National Laboratory (LLNL) examined the MHTGR as a plutonium burner and considered burnup cycles of 356,000, 560,000, and 730,000 MWd/MT plutonium, provided by General Atomics. Lack of information on the control characteristics for the pure plutonium particle cores led LLNL to examine a modified version, which would contain a neutron absorber or fertile material in the core. According to LLNL, this design would be within the current commercial design and safety envelope. LLNL concluded that this design could be deployed before 2010 if a decision was made in 1995, since the design is at a more advanced stage of development than the MHTGR plutonium burner sponsored by General Atomics. LLNL estimated the MHTGRs proposed by General Atomics, using a pure plutonium kernel and high burnup, would take an additional 5-10 years (Omberg and Walter 1993).
The principal advantage of the new PC-MHR approach, if the design holds up under detailed analysis (including safety reviews), is the possibility of lowering the cost of the plant while increasing the output, thus lowering the cost of electricity generated by the plant. While important commercially, this feature does not affect the plutonium-consumption characteristics. The MHTGR itself still has substantial licensing review ahead. This direct cycle system has much more.
Basic problems with the MHR include getting the funding for a new reactor (a problem common for any of the basic reactor approaches); developing and testing a new plutonium fuel; redesigning the core for higher burnup, in particular, if the fuel shuffling approach is to be used; and building and licensing a plutonium fuel fabrication facility. In addition, although the designs are different, any HTGR program will have to overcome utility concerns relating to the poor operational performance of the Fort St. Vrain reactor in Colorado, and the THTR (thorium high-temperature reactor) in Germany.
The reference PC-MHR facility has 14 modules, each at 600 MWt (286 MWe).42 Analysis has been done on several burnup programs. The reference design with a two-year fuel exposure is estimated by General Atomics to achieve 90-percent burnup of Pu-239 and 63-percent burnup of initial total plutonium, with an average burnup of 590,000 MWd/MT plutonium (GA 1994). At
590,000 MWd/MT, using fourteen 600-MWt MHR modules, 50 tons of plutonium could be processed to 90-percent Pu-239 destruction in 25 years from project start.
The MHTGR regular fuel cycle is 24 months long, with half of the core being replaced every 12 months. To achieve higher burnup, another proposal in the General Atomics Phase I report was to move fuel blocks to the reflector area after 24 months, for an additional 12 months of exposure. Shuffling fuel blocks to reflector locations for an additional one year of irradiation was reported to lead to 97 percent burnup of Pu-239 and 73 percent burnup of total plutonium, at 677,000 MWd/MT. Similarly, the Phase II report includes a brief mention of an alternative that would have a "slightly lengthened" refueling interval and would obtain 95-percent Pu-239 destruction and 72-percent total plutonium destruction (GA 1994, p. 1-20).
These once-through burnups are substantially higher than those projected for the LWR, CANDU, or ALMR systems because of the unique pellet design of the MHTGR and use of nonfertile fuel. Nevertheless, substantial quantities of plutonium would remain in the spent fuel (see below for a discussion of the difficulty of recovering this material for use in weapons). These burnups are therefore more than is needed for the "spent fuel" option, but less than needed for a true "elimination" option, which would require reprocessing and recycle.
In its Phase I report, General Atomics examined the concept of using two cycles of reprocessing and reuse of the plutonium to produce a higher final burnup. In this General Atomics concept, 80 percent of the fuel blocks loaded each year would be fresh blocks, 15 percent of the blocks would have been recycled once, and 5 percent twice. Using a simple model, this three-pass program is estimated to achieve 99.9-percent Pu-239 burnup and 90.7-percent total plutonium burning (GA 1993b, pp. 7, 16). At the end of a six-year residence time, the average fuel burnup would be 813,000 MWd/MT plutonium (INEL 1993a, p. 23). Additional recycling would be needed to eliminate the remaining material.
LLNL estimates it would take 10-15 years for deployment of a plutonium burner, based on the commercial variant using plutonium-uranium fuel, and 15-20 years for the plutonium fuel variants. The Idaho National Engineering Laboratory estimates 10-20 years for an MHTGR deployment after decision. Given such times, the MHTGR is not competitive with current-reactor designs for the "spiking" mission, with its emphasis on irradiating the WPu as quickly as possible, nor with the utilization of current LWRs for the "spent fuel" mission.
As noted above, fabrication and performance of plutonium HTGR fuels are among the biggest technical hurdles to be overcome for a plutonium-burning HTGR. Unlike the previous reactors discussed, in this case there are no substantial facilities anywhere in the world with the capability to undertake this fabri-
cation on the scale required. Licensing and construction of such a facility would be expensive and time-consuming.
Approvals and Licenses
The basic MHTGR design has been available for many years, although it has not been commercially successful. The MHTGR that had been proposed both for a commercial reactor and for the New Production Reactor had been under design for at least five years, although a completed design had not been developed. The MHTGR had been undergoing licensing review at the NRC,43 and the NRC has adequate experience for licensing review based on its work with Fort St. Vrain in Colorado.
The sponsor's principal concern for the commercial reactor appears to have been to make the case that no containment would be needed, because the pellets themselves would provide the primary containment, and therefore no credible scenario exists for release of radioactivity from the fuel. If this argument were accepted, the MHTGR would gain two principal advantages: cost and siting. A low-pressure confinement structure can be significantly less expensive than a full containment. And if the NRC agreed that there were no credible radiation release scenarios, the vendor argues that the NRC should drop the requirement for emergency planning and should allow the reactor to be sited close to industrial areas. Siting in industrial areas is a particular advantage for the MHTGR, which is a high-temperature reactor whose 700° C exit temperature (National Research Council 1992, p. 1 19) could be used to provide process heat.
Plutonium-kernel fuel will require extensive testing for licensing approval if the pellet integrity is to be relied on for a major element of the safety analysis. (As noted above, however, the plutonium-burner design is provided with a high-pressure, low-leakage containment, unlike the commercial design.) In addition, the direct cycle concept is new, leading to additional licensing issues. General Atomics notes (GA 1994, p. 13-1): "because the MHR is significantly different in technology and design philosophy from that with which the regulatory community is familiar, extra effort will be required . . . to familiarize the reviewers with the technology."
Safeguards, Security, and Recoverability
The MHR has a few unique features that affect the safeguards issue. The fuel form itself provides a barrier against ready access to weapon-usable plutonium. Once the plutonium has been placed in the kernel of the MHR fuel pellets, the silicon carbide coating makes recovery of the plutonium more difficult
than from normal fuel. Thus, manufacturing MHR fuel as soon as possible, well before it could actually be used in a reactor, would provide a more significant safeguard than would fuel fabrication for most other reactors.44 Overall safeguards risks during the fuel fabrication process would probably be comparable to those for other reactor types. Security risks in transportation of the fabricated fuel would be lower because of the greater difficulty of retrieving the WPu from the fresh MHTGR fuel.
Because of the higher burnup possible in an MHR, and the unique fuel characteristics, MHR spent fuel would pose a less attractive target for a potential proliferator than most other types of spent fuel. The argument should not be overstated, however. Once the WPu has been converted to spent fuel—and thus becomes only one small part of the much larger global stock of plutonium in civilian spent fuel—whether that small part of the total is significantly more proliferation-resistant or not, does not have a major bearing on overall safeguards and security concerns. That is the rationale for the “spent fuel standard."
In particular, while an MHR could potentially destroy a large fraction of the plutonium and leave the residual with high proportions of the isotopes above Pu-239, this does not greatly change the security picture when this spent fuel's place in the global stock is considered: whether the amount added to a global stock of 1,000-2,000 tons of plutonium in spent fuel is 50 tons or only 5 tons does not make a dramatic difference. And while the plutonium from high burnup MHR fuel would be more troublesome to use in explosives than standard reactor-grade plutonium, the fact is that nuclear explosives can be constructed from virtually any combination of plutonium isotopes (other than nearly pure Pu-238; see Chapter 2 and NAS 1994).
The volume and mass of the MHR fuel that would have to be stolen to acquire enough plutonium for a weapon is about 10 times larger than the equivalent volume for LWR fuel. On the other hand, because of the absorptive properties and small size of the graphite blocks, the radiation dose rate from MHR spent fuel would be substantially lower than the dose rate from standard burnup LWR fuel of equivalent age,45 making the MHR fuel somewhat easier to handle.
Separating the remaining plutonium from the MHR spent fuel would be more difficult than in the LWR case. The several layers of coating around each tiny particle would pose some barrier to recovery of weapons material. Methods to reprocess MHR fuel have been proposed, but there is no worldwide experi-
Note, however, that in the case of the ALMR, proponents have suggested a fuel form in which the WPu would be mixed with some minor actinides and fission products even before irradiation in the reactor, offering a significant safeguard against recovery of the plutonium. This concept is addressed in Chapter 5.
Ten years after discharge, the gamma dose one meter from the surface of a 100-kg MHTGR WPu fuel block that had been irradiated to 580,000 MWd/MTHM would be about 180 rem/hr, compared to about 940 rem/hr 1 meter from the surface of a 660-kg PWR WPu-MOX fuel assembly that had been irradiated to 40,000 MWd/MTHM (see Table 6-5).
ence with actual implementation of these methods comparable to the experience with reprocessing of LWR or LMR fuels. (This lack of experience is desirable for the proliferation resistance of once-through fuel, but undesirable if an elimination approach involving reprocessing of this fuel is to be pursued, as described above.) Since the kernels begin as pure plutonium oxide, however, there would be more plutonium per kilogram of initial heavy metal—once the kernel material was separated from the particles in the HTGR waste—than there would be in the spent fuel from other reactors.
The MHTGR is one of a set of advanced reactors designed to overcome the current utility- and public-acceptance problems seen as preventing expansion of nuclear power in the United States. Estimates of comparison costs indicate, however, that the MHTGR capital costs would be approximately 30 percent higher than those for ALWRs (National Research Council 1992, p. 139). Cost estimates for any unbuilt reactor design are uncertain. Cost estimates for development and fabrication of the plutonium fuel are even more uncertain. Nevertheless, numerous studies have concluded that the electricity from an MHTGR would be more expensive than that from ALWRs.
DOE concluded in Phase I of its Plutonium Disposition Study that, of the reactor options examined, only the HTGR did not produce a positive net value over its life cycle, using DOE's economic assumptions (USDOE 1993, Vol. 1, p. 5). General Atomics' calculations are more favorable to the MHTGR in some cases and less so in others; once discounting is applied, the General Atomics calculations also would not produce a positive net present value for the plutonium disposition campaign.
It is to address this cost differential that General Atomics has proposed the GT-MHR. The direct cycle is estimated by the vendor to lead to a 25 percent increase in thermal efficiency, as well as reducing the capital cost of the reactor. The vendor also proposes to increase the reactor to 600 MWt, also reducing the cost per megawatt. Further development and analysis will be necessary to validate the approach and cost improvements projected by the vendor.
There is no current infrastructure remaining for MHTGRs. General Atomics has maintained a cadre of technical staff who could begin design and testing immediately, however.
Environment, Safety, and Health
The MHTGR offers a potential advantage of improved safety against accidents compared to LWRs. If "containment-in-a-pellet" passes licensing approval, the reactor would be described as having an inherent safety substantially greater than LWRs. Such proof would require demonstrating, however, that the
number of fuel pellets that could deteriorate and provide a channel for radioactive product release would be insignificant.
The used fuel is taken out of the reactor with the fuel remaining in the graphite elements, thereby producing a large volume of high-level waste. Although it would be possible, in principle, to extract the fuel compacts, Oak Ridge National Laboratory (ORNL) has examined potential waste-disposal options for the MHTGR and has concluded that the preferred option would be disposal of the spent fuel still embedded in the graphite blocks. ORNL believes the fuel blocks could be placed in spent fuel waste containers similar to those for LWR spent fuel. General Atomics states that HTGR spent fuel could be packed more closely than LWR spent fuel because the HTGR waste has lower heat generation per unit volume, and that therefore the necessary repository size should not be increased significantly (see GA 1994, p. 13-2, referencing Lotts et al. 1992; GA 1993b, p. 2-335). DOE stresses, however, that the HTGR waste volumes are expected to be much greater than for the other plutonium disposition concepts. Current repository regulations bar emplacement of potentially flammable materials, including graphite. General Atomics argues that flammability is not a problem for nuclear-grade graphite (GA 1994, pp. 3-28 to 3-29).
Construction of a new MHR for burning WPu would be comparable to construction of a new LWR for the same purpose in terms of the signal relating to fuel-cycle policy that some parties might perceive such an action to send (see discussion in Chapter 6, "General Considerations").
Description of Technology and Status
The molten-salt reactor (MSR) would use a liquid fuel to carry the plutonium or other fissionable material into the reactor and to carry both heat and fission products out. The proposed approach has a spherical reactor vessel. The liquid fuel is a mixture of lithium, beryllium, and plutonium fluorides (possibly including some zirconium fluorides as well), which only attains criticality in the graphite moderator of the reactor core. Criticality excursions are limited by the large negative temperature coefficient of the fuel, which means that as criticality increases and more heat is produced, the reaction immediately slows down. The fuel circulates to an external heat exchanger. The heat-exchanger loop, which also uses a salt, goes to a steam generator for power production. A side stream is drawn from the reactor cooling loop, processed to remove fission products, and then returned to the reactor. In addition, the fuel salt would be in contact with helium for removal of volatile fission products (Gat undated). By this method of
continuously recycling the fuel, an MSR could, in principle, achieve near-total elimination of plutonium added to the system.
As an advanced reactor requiring a long period of development and demonstration before a commercial-scale system could be built, the MSR is not competitive for the spiking or spent fuel options, so these are not discussed further in this section. It could be a candidate, however, for the elimination mission.
Liquid-fuel reactors have two main advantages: ease of heat removal and fuel management. They also have the potential for achieving high power density. The MSR began as part of a program to develop a nuclear-powered aircraft. The reactor development effort was shifted to a commercial power design in 1956, and to a breeder program in 1960 (McPherson 1985). In the early 1950s, the liquid-fuel reactor design envisioned uranium fuel dissolved in a molten salt. The early view of the utility of molten salts was mixed (Glasstone 1955, p. 527; Lane et al. 1958). Supporters of molten-salt systems (located primarily at the Oak Ridge National Laboratory) believed, however, that the Atomic Energy Commission decision to concentrate on the LWR and the sodium-cooled reactor was a mistake, and they have maintained interest in the MSR ever since, although basic funding for the program ended in the mid-1970s. Because development of this concept was abandoned two decades ago, the available base of expertise in molten-salt concepts is minimal. Recent interest in molten-salt systems has focused primarily on subcritical reactors driven by accelerators, an approach known as accelerator-based conversion (ABC, described below), though some study is now being devoted to molten-salt systems without accelerators.
Current MSR concepts are based on the Molten-Salt Reactor Experiment (MSRE), a low-power reactor (8 MWt), which operated with various salt compositions that included both uranium and plutonium.46 The MSRE was initially operated with U-235 at 35-percent enrichment and operated for nearly three years, from 1965-1968. A U-233 fuel was added after removal of the U-235. The plutonium produced remained in the salt and several additions of PuF3 were made, though uranium remained the dominant fuel. Later work concentrated on developing an MSR thorium breeder, using a 7LiF-BeF2-ThF4-UF4 salt.
Several problems were identified during the operation of the MSRE and in subsequent analysis before the program was terminated. These included developing graphite with higher radiation resistance and lower permeability to the salt and to xenon produced during fission, so that the graphite would not have to be replaced every few years. A second problem was that the creep-ductility of the material that had been used for the vessel and piping, Hastelloy N (a nickel-base
alloy), was reduced by neutron radiation. A solution under investigation at program termination was to add 2 percent titanium. An environmental problem is the production of tritium, produced in lithium in the fuel salt; tritium diffuses readily through metals at the MSRE operating temperatures (McNeese and Rosenthal 1974, pp. 55, 57-58). Use of a mixture of sodium, fluoride, and sodium fluoroborate to capture the tritium had been proposed as a solution (McPherson 1985, p. 377). Further discussion of molten-salt issues can be found below in the section "Accelerator-Based Conversion of Plutonium."
Compared to other systems discussed in this chapter, little information is available on the MSR. The MSR sponsor recently wrote: "there is no active program on molten salt reactors and, hence, we cannot make any additional calculations, no matter how simple they are" (Gat 1993, p. 1). The sponsor noted that "there is no current cost estimate for fuel processing" and “the estimated duration of [full] development has such wide margins that any number or even range is meaningless."
Reprocessing in the MSR concept is part of the integral fuel management. A fuel loop would carry a portion of the fuel continuously through a reprocessing station, where fission products would be removed and the actinides and salt would be returned to the reactor. The fissionable material would be recycled until consumed. The basic processing steps were demonstrated by the MSRE, with the important exception of demonstrating how to separate plutonium from the salt in the presence of uranium and fission products, which would require development effort (Omberg and Walter 1993, p. 18).
MSR advocates have always stressed potential safety features of a molten-salt reactor, including possible passive safety. Proponents argue that the MSR design has a low source term due to its small radioactive inventory and a lack of energy to drive an offsite release, and that its simplicity and passive safety features add to its overall safety (Gat 1986). Some safety issues remain unresolved, however, including possible reactivity excursions, containment of radioactive tritium (although the intermediate-loop proposal may solve that problem), and the containment of the hot, radioactive liquid fuel in the case of a pipe break.
Timing and Other Issues
The MSR exists only as a conceptual design. Considerable development effort would be needed before any MSR system could be fielded. A study by the Lawrence Livermore Laboratory estimated that 23-27 years would be required for deployment after decision (Omberg and Walter 1993, p. 28). No calculations have been done concerning how long it would take to eliminate specific fractions of the input plutonium (e.g., 99 or 99.9 percent) in an MSR system.
The sponsor of the MSR has noted the licensing problems associated with such systems, arguing that "perhaps the biggest availability issue is the licensing of MSRs. Present licensing practices are oriented solely toward LWRs and solid
fuel systems. . . . The licensing of MSRs, keeping their advantageous property intact[,] by NRC will require a large effort" (Gat undated, p. 8). Given the novelty of the reactor concept, public acceptance is an open question, and it is highly unlikely that any utility would be interested in trying this reactor.
The economics of an MSR system are too uncertain to estimate with any confidence. The long development effort required to bring the MSR to the level of technical maturity of the ALWR, ALMR, or MHTGR would be costly, however; and it appears unlikely that an advanced fluid fuel reactor system with reprocessing can be competitive economically with LWRs in producing electrical power while uranium prices remain low.
Estimating the ES&H impacts of an MSR system is nearly impossible at this early stage. The MSR may offer a reduced risk of large accidents, but also may have a larger number of scenarios that could lead to radioactive contamination within the reactor building, leading to greater worker hazard. The intensely radioactive molten fuel Would lead to high radioactivity levels throughout the reactor primary circuit, requiring development of remote-handling equipment and potentially posing risks of significant worker exposures.
Safeguards and security risks for an MSR system should be low, once the plutonium is mixed with the radioactive molten fuel. If the plutonium is recycled until it is nearly all consumed, the risk of recovery after disposition would be effectively eliminated.
Description of Technology and Status
The particle-bed reactor (PBR) concept evolved from designs for nuclear powered rockets for the Strategic Defense Initiative program, a concept codenamed Timberwind. The PBR fuel elements consist of two concentric cylinders with fuel in tiny particles packed between them. The fuel particles (diameter 0.8 mm) would be similar to those of the MHTGR. The PBR fuel particles have a central graphite kernel that, in the version under development, contains uranium carbide. The plutonium burner would have a kernel of a graphite matrix with PuC2, a graphite layer, and an outer coating of pyrolytic graphite. If fuel tests indicate it is necessary, a final layer of silicon carbide would be added (Brookhaven 1992, pp. 2, 40; Ludewig 1993b).
The inner and outer cylinders are constructed of porous tubes called frits. Gas coolant flows through the center core and outwards through the cool inner frit, across the fuel pellets, and then through the hot outer frit. Fuel elements are arranged in hexagonal patterns and surrounded by a moderator, vital to the operation of the system, which could be graphite, beryllium carbide, heavy water, or other materials. The proposed design uses helium coolant, with the cold frit at approximately 300° K and the hot frit at about 1,000° K. The hot frit would be
made of Incalloy and the cold frit made of Zircalloy. Very high power densities are estimated for this design—approximately 5 megawatts per liter.47 Therefore the reliability of coolant flow must be high.
The basic PBR design, developed for nuclear rockets, has been discussed for many years. The proponent of this concept is Brookhaven National Laboratory (BNL), whose work has been done for the Air Force. The space nuclear propulsion program has been terminated, however, and the BNL work on it ended at the end of fiscal year 1993. A low level of effort continued, aimed at developing proposals to use the PBR for plutonium and actinide burning (Ludewig 1993c).
The BNL program has developed the uranium fuel particles, the frits, and control designs, and has analyzed the thermal hydraulics and neutronics. The design also has been proposed as a means of transmuting actinides and long-lived fission products from LWR spent fuel. In that case, target elements containing fission products and minor actinides would be introduced along with fuel elements.
The reactor would be shut down during loading and unloading operations, estimated to be required at least every two to four weeks.48 The proposed design would have a hydraulic loading and unloading operation, a concept that has not been tested on PBR nuclear fuel elements.
Since the design was being developed for a rocket, no significant effort had been devoted to considering how the design could be used as a power reactor. BNL estimates that a conventional steam cycle based on MHTGR experience would be used. The use of helium as a coolant has been tested at low pressure and with electrically heated particle beds. The design uses beryllium carbide, Be2C, in a graphite structure as the moderator, which would require developing manufacturing techniques and raises some potential ES&H issues. The basic fuel particle, PuC 2, has not been fabricated or tested (Koopman et al. 1992, p. 4-36).
Potential problems that have arisen in testing include the following (Brookhaven 1992, pp. H2-H5; INEL 1993a, pp. 10-1 1):
Chemical compatibility of the fuel, coatings, fission products, and coolant.
Liquefaction of the hot frit, either because temperatures were hot enough to melt the hot frit or because of thermal interaction between the hot frit and the fuel.
Significant fuel swelling due to retention of gaseous fission products, which has reached 20-30 percent and could lead to extensive fuel failure.
Potential complications in fuel fabrication from impurities in the plutonium. Extrapolation from the uranium fuel fabrication technology to plutonium is not warranted.
Shortening of the hot frit with cycling, which can lead to fracture and loss of fuel pellets. Nuclear testing has resulted in hot frit shortening and numerous cracks in one element, which may limit the number of times the PBR fuel elements can be thermally cycled.
Thermal and hydraulic flow stability in the particle fuel bed.
Timing and Other Issues
As with the MSR, the recent LLNL study concludes that deployment of a commercial-scale PBR would require at least 20-25 years after a decision was taken (Omberg and Walter 1993, p. xiii). Hence the PBR also would not be competitive with other approaches for the spiking or spent fuel missions.
Because of the extremely high power density in the PBR, consumption of actinides in the system would be rapid. BNL estimates that the PBR could reduce a small initial plutonium inventory by 95 percent in 20 days, based on a burnup of 500,000 MWd/MTHM (Ludewig 1993a, p. 2). Reprocessing would then be required if a true elimination option was to be pursued. BNL states that techniques exist to reprocess the spent fuel particles for this purpose. As in the case of the HTGR, however, reprocessing approaches for this fuel require considerable development, as would fabrication technologies for multiple recycle materials.49
As with the MSR, licensing and public acceptance of an advanced and unfamiliar reactor type would be time-consuming and uncertain. The NRC has not licensed a similar type reactor. Many questions would need to be answered before a licensing estimate could be made. Serious problems are to be expected.
ES&H impacts of a PBR system are also uncertain. There would be a low volume of radioactive waste in the fuel, but the reactor components may need to be discarded on a much more frequent cycle than that for current-reactors, leading to a higher total volume of radioactive waste. Current repository rules bar emplacement of potentially flammable materials, including graphite; as with the MHTGR, this problem would have to be addressed. Frequent refueling would increase the risk of worker exposures. The design has not been developed far enough to provide an estimate of safety. Because of the small core, the total amount of fission products contained at any one time is small compared to a
normal LWR. In that sense, the hazard may be lower. Nevertheless, given the high power density, careful design will be required to provide adequate emergency cooling to prevent excessive heating of core materials in the event of a loss-of-coolant accident. Moreover, because of the high burnups projected on a single pass, reactivity swings in the reactor will be substantial. Uncertainty masks any safety estimate.
Similarly, it is too early to guess what the cost of a PBR system would be. Given the early stage of development, and the uniqueness of the concept, developing and deploying a PBR for WPu disposition would probably be significantly more expensive than developing and deploying any of the concepts already well into the development cycle.
A DEDICATED PLUTONIUM-BURNER REACTOR
To ensure that the full range of possibilities was appropriately explored, the panel considered what a dedicated plutonium disposition reactor designed for no other purpose would look like. The mission of such a dedicated disposition reactor would be to transform plutonium into spent fuel as quickly, safely, and cheaply as possible—without the accompanying mission of producing electricity in the process.
Removing the mission of producing electricity would simplify the design of a reactor substantially. There would be no need for high temperature to operate a heat engine, or for the high pressure that accompanies high temperature in the case of water coolants. Capital costs would be much reduced by eliminating turbines, generators, many of the support facilities, and the thick-walled pressure vessel needed to contain a high-pressure system. Reactor design would be simpler and more flexible. Operation at low temperature and pressure would minimize the required strength of reactor vessels, piping, and the like. In fact, the reactor could operate in a warm pool of water and rely on natural circulation for removal of decay heat. The fuel temperature would also be low, allowing one to use fuel that could be more cheaply fabricated (an important part of the cost of operations in a plutonium-fueled reactor). Safety could be significantly increased, in part because the low operating temperature would result in little or no steam energy with which to deal in the case of an accident.
At the panel's request, the Idaho National Engineering Laboratory (INEL) outlined a conceptual design of such a reactor (INEL 1993b). They envisioned a reactor with a core of comparable size to those of modern LWRs, but with a power density 10 times lower, and a maximum fuel temperature in the range of 400° K, rather than over 2,000° K in the case of a PWR. The core envisioned would have a thermal power of 1,000 megawatts.
A fuel with particularly low-cost fabrication would be plutonium-aluminum alloy extruded in aluminum cladding. Since the goal would be to secure plutonium by getting it into the reactor as quickly and cheaply as possible, fuel with
very high plutonium loadings would be highly desirable. This could be achieved with the use of large quantities of burble absorbers, such as erbium, to offset the extra reactivity. It would be quite feasible to design such a core to be loaded with 10 tons of plutonium at a time.50 This would mean, however, that when discharged, the fuel would be substantially richer in the remaining plutonium than ordinary LWR fuel (by a factor of 10). This would have to be balanced against the reduced fabrication cost. Detailed study of whether the cheaply produced aluminum-clad fuel would be an acceptable waste form for geologic disposal would also be required. Such fuels have historically suffered corrosion problems in storage, leading to their being reprocessed rather than stored in most cases, and certifying such a fuel type for repository disposal would be difficult, time-consuming, and costly. Stainless steel cladding is an alternative proposal that would also have relatively low cost.
All told, the capital cost of such a reactor would be expected to be hundreds of millions of dollars less—perhaps more than a billion less—than that of a typical LWR, and operations costs would also be expected to be less.51 But since this reduction in cost would come with the sacrifice of many billions of dollars in revenue that an electricity-producing reactor would provide, it appears extremely unlikely that the net discounted present cost of developing, licensing, building, and operating such a system for plutonium disposition would be competitive with the costs of using MOX in existing or new LWRs. Design, development, and licensing of the reactor and its fuel in the current regulatory environment in the United States would be expected to take many years, moreover. Thus the panel does not recommend further pursuit of the option of a specially designed plutonium burner without electricity production.
ACCELERATOR-BASED CONVERSION OF PLUTONIUM
Description of Technology and Status
Accelerator-based conversion (ABC) systems have been under study as a means of eliminating plutonium and of fissioning actinides and transmuting fission products in order to reduce the longevity of radioactive wastes. Despite the name, these concepts do not involve bombarding the plutonium directly with an accelerator. Rather, these concepts are a variation on reactor burning of plu-
tonium: a subcritical reactor—meaning one that is not capable of sustaining a chain reaction without outside neutron input—would be driven by neutrons produced by a beam of particles from an accelerator hitting a target. Typically the proposed reactors would have a multiplication factor, k, of about 0.95, meaning that 19 out of every 20 neutrons in the system would come from fission, not from the accelerator. A portion of the power provided by the reactor (typically 10-20 percent) would be used to power the accelerator.
The accelerator would act, in effect, as a negative control rod: shutting off the accelerator would, if everything worked as intended, bring the reactor below criticality and shut down the reaction. This could be done more rapidly than mechanical control rods can be inserted in an ordinary reactor, which advocates argue would lead to improved safety—particularly in the case of plutonium fuels, with their smaller delayed-neutron fraction. Proposed ABC systems, however, raise some reactor safety issues of their own, which have not been fully resolved.
Analysis of ABC systems is hampered by the existence of a wide array of fast-changing proposals. Virtually any type of reactor could be arranged in a subcritical state, to be driven to criticality by spallation neutrons from an accelerator hitting a target.52 Liquid-fueled reactors with both aqueous-slurry and molten-salt fuels have been examined, as has a gas-cooled particle-bed system. LWRs, LMRs, or HTGRs could also be considered. During 1994, the approach being most energetically examined was a molten-salt system; this system is the focus of discussion in this section. Just as the reactor subsystem is uncertain, the power level of the accelerator has not yet been optimized, nor has the spallation target that would produce the neutrons when bombarded by the accelerator been fully developed.
ABC concepts are only at the early paper-study stage, and such devices could not be operational on a large scale for decades. Both the proposed fluid-fuel subcritical-reactor technology and its fuel-cycle technology are unproven and extremely challenging. As such, these advanced concepts are a potential long-term option for the elimination mission, but are not competitive with nearer-term reactor types for the spiking or spent fuel missions.
Nevertheless, ABC research efforts at Los Alamos have shifted, at DOE direction, from an initial focus on developing an option for nearly complete elimination of both WPu and the much larger global stocks of RPu, to a current focus on a once-through approach for burning of WPu, without explicit consideration of recycling. Under this approach, a significant fraction of the plutonium would
remain in the spent fuel. A recent review has concluded that "this narrowing of mission" from the total world plutonium stock to burning a fraction of the WPu stock is "unwise," and that "the ABC concept as refocussed under guidance is not worthy of support” (JASON 1994, pp. 17, 18). In this section, therefore, we examine an ABC refocused on the broader elimination mission, which would require reprocessing and recycle. As argued elsewhere in this report, incurring substantial additional delays, costs, or risks to pursue the elimination option for WPu should only be considered if the much larger global stocks of plutonium in the civilian cycle are also to be consumed.53
In a number of the proposed ABC approaches involving reprocessing, both actinides and long-lived fission products would be separated and recycled. Excess neutrons, either from the subcritical reactor itself or from a separate accelerator system, could be used to transmute the long-lived fission products (particularly technetium-99 and iodine-129) into short-lived species.
ABC advocates claim that such systems will ultimately be able to transmute all the long-lived species produced in the reactor, providing large-scale nuclear electricity generation with virtually no long-lived radioactive wastes, while simultaneously transmuting the wastes from previous generations of nuclear power. In principle, resource extension could also be achieved, through use of a thorium-uranium breeder cycle. This concept of vast resources of virtually waste-free power, however, remains little more than a vision at present. A wide array of key technologies involved remain to be demonstrated.
In particular, the panel notes that if the cost of power produced by such a system were even a few mils per kilowatt-hour higher than that of electricity produced by other means-which appears very likely, given that the system involves advanced reactors with extensive and complex reprocessing and with a significant fraction of the electricity produced being siphoned off to run the accelerator—then the total subsidy needed for transmutation of all the actinides and long-lived fission products in the spent fuel already produced in the world would amount to many hundreds of billions of dollars.
A Baseline System
As currently envisioned,54 a baseline ABC system would consist of the technologies discussed in the following paragraphs.
The accelerator power required depends on how far the reactor is from criticality (that is, the value of k), and this parameter has not yet been optimized. Similarly, the current and proton energy of the accelerator could cover a broad range. As of January 1994, the Los Alamos team advocating the ABC approach envisioned an accelerator with a beam energy of 800 million electron volts and a current of 75-100 milliamperes.
The accelerator beam would be split between four to eight target modules, each consisting of a beam target and a subcritical reactor. A variety of material choices are possible for the target of the proton beam; the target would be designed to produce of the order of 25 neutrons per proton.
These neutrons would go into a subcritical reactor system, which would consist of molten salts (lithium, beryllium, and zirconium fluorides), with plutonium dissolved in them, circulating within graphite blocks that would serve as the moderator. Each module might have a power of roughly 500 MWt, so that an eight-module system might have a net output power of 1 GWe (after subtracting 10-20 percent of the electrical power to run the accelerator). The neutron flux in these modules would be extremely high—initially 3-8 × 1014 n/cm 2 (neutrons per square centimeter). The reactor would be cooled by circulating the fuel through heat exchangers immersed in the molten salt. As envisioned in January 1994, each module designed for the WPu destruction mission would begin with roughly 20 kg of plutonium in the fuel. The fuel would continue to reside in the reactor throughout the 10-year expected module life; additional plutonium would be added over time to make up for the loss of reactivity resulting from plutonium consumption and the buildup of fission product poisons. All told, 2.3 tons of plutonium would be added to each module over its 10-year life, of which approximately 90 percent would be fissioned during that period.
Reprocessing and Enrichment
As noted above, in some current concepts there would be no reprocessing. The actinides and fission products that remained in the fuel at the end of the 10-year module life would simply be vitrified and disposed of in a geologic repository. This would be radically different from earlier ABC and accelerator transmutation of waste (ATW) approaches, which advocates claimed could minimize or eliminate long-lived wastes by reprocessing and recycling the actinides and long-lived fission products. In the no-reprocessing approach, there would still be significant quantities of plutonium in the waste. Thus this approach is another, and expensive, way to burn a fraction of the plutonium while
leaving the rest embedded in radioactive waste. It cannot be considered a genuine "elimination" option.
The fraction of the plutonium consumed would be higher than in the case of LWRs (and certainly higher than in the vitrification approach, where none of the plutonium is consumed), but this difference would offer little security advantage, since the plutonium in the waste, whether it was 5 tons or 50 tons, would be only a small part of the many hundreds of tons of plutonium in spent fuel worldwide. Given the long time, high cost, and high uncertainty that would be involved in ABC development, the panel does not believe that ABC would be competitive with other systems unless it were designed for the elimination mission, which would require reprocessing and recycle.
Reprocessing has been examined for these concepts and demonstrated on a limited scale in the MSRE. If both actinides and long-lived fission products are to be transmuted, these would both have to be separated from the short-lived fission products, which would be stored onsite until they decayed. Originally, advocates of fluid fuel ABC concepts proposed an extremely challenging online reprocessing approach in which small amounts of fluid fuel would be continuously removed from the reactor, reprocessed with very little delay (meaning that the fuel would be orders of magnitude more radioactive than solid fuels in past reprocessing experience), and the actinides, long-lived fission products, and molten salt would be returned to the reactor. The average time each atom spent in the reactor between reprocessing cycles would have been only a few weeks, and on average, each atom of fissile material would have gone through more than 100 reprocessing cycles before being fissioned. Recovery of the important radioactive species requiring further transmutation would have to be extremely high to avoid these species building up in the waste during these repeated reprocessing cycles. In addition to chemical separations, enrichment of cesium isotopes would be required, because of the need to transmute long-lived cesium-135 (Cs-135) without burdening the reactor with comparatively short-lived Cs-137. The challenging separations needed for this concept have not been demonstrated and are likely to pose larger technical difficulties than would the other reprocessing concepts discussed in earlier sections of this report.
In the more recent concept, reprocessing and enrichment would be required only after the fuel's 10-year residence in the reactor module. This would allow greater cooling time for the fuel before reprocessing. Still, the required molten-salt processing would pose substantial challenges, and the problem of enriching intensely radioactive cesium isotopes—which has never before been done on the required scale—would remain.
Because the plutonium would not be separated from the minor actinides and long-lived fission products in this reprocessing approach, the risk of plutonium theft or covert diversion would be substantially lower than in reprocessing concepts in which the plutonium is fully separated-an advantage that ABC shares,
to varying degrees, with other reactors using similar reprocessing approaches, such as the IFR.55
Balance of Plant
As with other reactors, the heat from the ABC would be used to drive turbines to produce electricity. As noted, 10-20 percent of this electricity would be used to power the accelerator.
Alternatively, the ABC could be supplied with external power for the accelerator and use a cool slurry rather than one which is hot enough to provide (through a heat exchanger) steam to drive a turbogenerator. In this way, the accelerator and subcritical reactor could be divorced from the cost and constraints involved in providing electrical power. Using ABC systems to transmute a large fraction of all the actinides in the world, however, would not be economically practical unless electricity were produced to cover at least a portion of the cost.
The accelerator part of the ABC system is close to the existing state-of-the-art, but every aspect of the reactor and reprocessing systems involves important unknowns.
The average power of the accelerator would be well beyond that of any accelerator previously built, but the peak power, particle energy, and current required have all been demonstrated. While the accelerator would require considerable development, there is no reason to believe that an accelerator with the required capabilities could not be built on a reasonable time scale. Similarly, technology appears to be available to build an accelerator target that would produce the required spallation neutrons.
The proposed fluid fuel subcritical reactors are based on immature technology that still faces major challenges. The reactor technology is based on the much smaller (8-MWt compared to 500-MWt) Molten-Salt Reactor Experiment (MSRE) which operated from 1965-1968 (see description in section on "The Molten-Salt Reactor" above). As noted earlier, the MSRE program was terminated in the 1970s, with other reactors chosen for development, meaning that
only a modest current knowledge base is available to answer questions regarding molten-salt systems.
The MSRE identified several major issues that were not resolved when the program terminated. In addition, several aspects of molten-salt ABC systems would differ substantially from the MSRE, posing new challenges.
Loss-of-Coolant Accidents. One of the most significant safety issues in power reactors is the need to deal with fission-product heating after shutdown. The proposed ABC system relies on heat exchangers for cooling. Systems will be needed to ensure adequate cooling in the event of heat-exchanger failure. Design for such an accident would require means for dumping the contents of the reactor into a volume that maintains the fuel in a subcritical geometry, prevents the intensely radioactive fission products from escaping or otherwise posing a safety hazard, and provides reliable, effective means for removing hundreds of thermal megawatts of fission product decay heat. Earlier aqueous-slurry ABC concepts would not have met the basic safety criterion of being able to survive a pipe-break accident, as such an accident would have led to intensely radioactive slurry fuel spraying out into the containment.
Reactivity Excursions. The choice of k = 0.95 reflects an implicit assumption that there will be no reactivity excursions of more than 5 percent. This assumption may or may not be correct and must be thoroughly demonstrated if the reactor is to be judged safe. Fluid fuel reactors, for example, tend to have very large negative temperature coefficients of reactivity (because of the expansion of the fluid fuel as it heats up). Therefore going to cold shutdown would greatly increase the reactivity of the reactor—possibly by an amount larger than the capacity of the accelerator to control. Control rods similar to those used in more conventional reactors may be needed for safe shutdown.
Another problem is potential xenon oscillations. High neutron-flux thermal reactors tend to experience spatial oscillations in power density, caused by xenon-135 produced during the reactions, which can cause local overheating. Xenon oscillations are a significant safety issue in LWRs and even more so in heavy-water moderated plutonium production reactors; they could be more severe in an ABC system, with its higher neutron flux. (Such fluctuations were not a major problem for the MSRE, because of its small size.) Moreover, they could be exacerbated by xenon encroachment in the graphite moderator, which was observed in the MSRE. (Some approaches to sealing the graphite appeared to limit xenon encroachment substantially in the MSRE experience, but graphite permeability increased under neutron bombardment; this problem was not solved before the program was terminated.)
A further problem arises from the very large neutron absorption cross section of samarium-149 (Sm-149), a daughter of the fission product promethium-149, which has a 53-hour half-life. Shutting down a high-flux reactor can
lead to the growth of such large amounts of Sm-149 that it may not be restartable until the Sm-149 has been removed.
If the ABC were partly fueled with thorium, as envisioned in some concepts, another potential problem is created by the isotope protactinium-233, which decays to U-233; when power is reduced from equilibrium, the amount of U-233 in the system will increase, causing an increase in reactivity. The higher the neutron flux, the larger this reactivity swing would be, raising a concern for the high-flux ABC.
In short, considerable work needs to be done before there can be high assurance that the risk of potentially dangerous reactivity excursions in the proposed subcritical reactors can be eliminated.
Use of Plutonium. One of the important differences between the proposed ABC molten-salt system and the MSRE is the use of plutonium as the fuel. Only a small amount of plutonium was ever introduced into the MSRE, which ran primarily on uranium. A technology development program would be needed to demonstrate that the different neutronic and chemical characteristics of the system with plutonium are adequately understood.
For example, fission of uranium tends to make the chemistry of the molten salts more oxidizing, while fission of plutonium tends to make it more reducing; a reducing environment could lead to precipitation of certain metals, which could have serious effects (such as blocking flow and thereby leading to local overheating).
Miscellaneous Issues. Other issues identified in the MSRE include surface cracks and ductility changes in the alloy used to contain the salt, graphite permeability to the salt, and shape stability of the graphite under intense irradiation. Warping of the graphite blocks in an ABC system could create voids where fuel might collect and stagnate, causing local overheating. The production of tritium from bombardment of the lithium in the salts is a potential ES&H concern, particularly given the permeability of the containment metals to tritium at the relevant temperatures. Similarly, the reliance on highly toxic beryllium fluoride in the MSRE might not be acceptable under today's ES&H standards.
The reprocessing involved in a molten-salt system, particularly one designed to transmute long-lived fission products as well as actinides, would be quite different from the aqueous reprocessing used around the world today. There is no base of experience available comparable to the experience with aqueous reprocessing. As noted above, enrichment of intensely radioactive cesium would also be required. To achieve the hoped-for goal of having the reprocessing waste qualify as Class C waste (not requiring repository disposal), unprecedented separation factors would be required. (It should also be noted
that even if this goal is met for the fuel itself, the irradiated system components and similar wastes will almost certainly require repository disposal.) The necessary processes require development and demonstration.
Considerable further study will be required to optimize the various components of the ABC concept. One important question in that process will be: Is the accelerator that gives the concept its name needed? An alternative would be operating a similar reactor system (whichever reactor system is ultimately chosen for ABC) at full criticality, as current reactors operate (and as the MSRE operated). Does subcritical operation provide a unique safety advantage sufficient to justify consuming 10-20 percent of all the power produced by the system (meaning 10-20 percent of all its potential revenues in electricity sales) to operate the accelerator? If an accelerator is needed, would a lower-power accelerator (and a reactor closer to criticality) be sufficient?
In short, the ABC technology remains quite immature. Technical feasibility of a number of the fundamental concepts, and engineering feasibility of essentially every aspect of the system, remain to be demonstrated. A development program costing several billion dollars and lasting for decades would probably be needed before a commercial-scale ABC system could be built with confidence.
The ABC concept is too immature to allow detailed treatment of issues such as licensing, ES&H, and cost, but some discussion is in order.
As with other fundamentally new reactor designs, gaining required licenses and political approvals for construction and operation of an ABC system is likely to be difficult. The NRC has no experience regulating systems comparable to proposed ABC molten-salt reactors. Both the new-concept reactors and the associated reprocessing may face substantial difficulties with public acceptance. Even if all long-lived wastes could be transmuted, the storage of tanks of intensely radioactive liquid wastes would be required at the ABC site for decades or centuries, a further complication for public acceptance. If, however, claims that the system can transmute all long-lived wastes are proven, prospects for acceptance could be improved.
ES&H impacts of ABC are claimed to be low. ABC proponents point in particular to the hoped-for nearly complete elimination of long-lived species that must be disposed of in geologic repositories, which would reduce the potential ES&H impacts of repository disposal. Success in a wide array of as-yet-unproven technologies will be required, however, to ensure that ES&H impacts of ABC-system operations (including reactor operations and reprocessing) will
be acceptably low. As noted earlier, while molten-salt systems may offer a reduced risk of large accidents, there may also be a larger number of scenarios that could lead to radioactive contamination within the reactor building, leading to greater worker hazard. The intensely radioactive molten fuel would lead to high radioactivity levels throughout the reactor primary circuit, requiring development of remote-handling equipment, and potentially posing risks of significant worker exposures.
The cost of ABC systems is quite speculative, given their current state of development. If past experience is a guide, costs will be substantially higher than now envisioned. It appears unlikely that an advanced-reactor system requiring a large accelerator and advanced reprocessing will be cost-competitive with proven LWR technology as a power producer in the next several decades, while abundant low-cost uranium continues to be available. If ABC were not cost-competitive, so that a "transmutation subsidy" of, for example, a few mils per kilowatt-hour would have to be paid to make the system economically viable, the cost to transmute the world supply of spent fuel could be very large.
If ABC existed and were in use for transmutation of high-level wastes and actinides, then it could readily handle the WPu. The development and validation program for ABC would take longer than for most other options, however, and it is impossible as yet to have high confidence that this approach will succeed.
The role of the ABC is far from assured. First, it must be proved feasible—technically, politically, economically, and institutionally. If it is feasible, there seems no reason why its rare abilities should be wasted on the relatively easy task of converting WPu into spent fuel. Accelerator-based conversion might be of use after the WPu problem has been transformed into a small part of the civilian plutonium stock. Only if that large global stockpile were to be transmuted by ABC systems would it make sense to commit the WPu to treatment by accelerator-based conversion.
ABB-CE 1989: ABB-CE Combustion Engineering. "SIR Design Summary." Presentation to Nuclear Regulatory Commission, Rockville, December 1989.
ABB-CE 1993: ABB-CE Combustion Engineering. Plutonium Consumption in ALWRs. DOE Plutonium Disposition Study, Final Report. Windsor, Conn.: ABB-Combustion Engineering, May 15, 1993.
ABB-CE 1994: ABB-CE Combustion Engineering. DOE Plutonium Disposition Study: Analysis of Existing ABB-CE Light-Water Reactors For The Disposition of Weapons-Grade Plutonium . Contract DE-AC03-93SF19682. Windsor, Conn.: ABB-Combustion Engineering, June 1, 1994.
AECL 1994: AECL Technologies, Limited. Plutonium Consumption Program CANDU Reactor Project. Contract DE-AC03-94SF20218. AECL Technologies, July 31, 1994.
Albright et al. 1993: David Albright, Frans Berkhout, and William Walker. World Inventory of Plutonium and Highly Enriched Uranium 1992. London: Oxford University Press for the Stockholm International Peace Research Institute, 1993.
Anderson and Klinetob 1981: G.K. Anderson, and R.H. Klinetob. "Plutonium-Burning Light-Water Reactor Concept." Combustion Engineering, TIS7016, September 1981.
Baumgartl and Watteau 1994: Bruno J. Baumgartl and Michel P. Watteau, Nuclear Power International. "Current Status of the Joint French-German Development of the Next PWR Generation." Paper presented at International Topical Meeting on Advanced Reactors Safety, Pittsburgh, April 1994.
Berkhout et al. 1993: Frans Berkhout, Anatoli Diakov, Harold Feiveson, Helen Hunt, Edwin Lyman, Marvin Miller, and Frank von Hippel. "Disposition of Separated Plutonium." Science and Global Security 3:161-213, 1993.
Bibilashvili and Reshetnikov 1993: Yu. K. Bibilashvili and F. G. Reshetnikov. "Russia's Nuclear Fuel Cycle: An Industrial Perspective." IAEA Bulletin 3/1993.
Blue 1993: Linden Blue. "New Technologies Compel the Combination of Gas Turbine with Nuclear Energy." Briefing enclosed in letter from Linden Blue, General Atomics, to John Ahearne, May 7, 1993.
Boczar et al. 1994: P.G. Boczar, S. Kupca, P.J. Fehrenbach, and A.R. Dastur. "Plutonium Burning in CANDU." Paper presented at NATO Advanced Research Workshop on "Mixed Oxide Fuel (MOX) Exploitation and Destruction in Power Reactors," Obninsk, Russia, October 16-19, 1994.
Brookhaven 1992: Brookhaven National Laboratory. "Section 3.3 × Description of the Particle Bed Reactor Plutonium Burner." Submitted to U.S. Department of Energy Plutonium Destruction Study. Brookhaven National Laboratory, 1992.
Bruschi and Andersen 1991: Howard Bruschi and Theodore Andersen. "Turning the Key." Nuclear Engineering International November 1991.
Carter 1977: Jimmy Carter. Presidential Documents—Jimmy Carter. 13(15), April 18, 1977.
Choi and Pigford 1993: J.S. Choi and T.H. Pigford. "Reduction in Transuranic Inventory by Transmutation." UCB-NE-4183 (Rev. 1), 1993.
Ebbinghaus et al. 1992: K. Ebbinghaus, M. Debru, and C.H. Mitchell. "Technical Progress with the EFR Design." Paper presented at the International Conference on Design and Safety of Nuclear Power Plants, Tokyo, October 1992.
EBRD 1993: European Bank for Reconstruction and Development. Nuclear Safety Account for Upgrading Soviet-Designed Reactors. London: EBRD, 1993.
Egorov et al. 1994: N.N. Egorov, V.M. Murogov, V.S. Kagramanian, M.F. Troyanov, V.M. Poplavski, and V.I. Matveev . "Management of Plutonium In Russia." Paper presented at NATO Advanced Research Workshop on "Mixed Oxide Fuel (MOX) Exploitation and Destruction in Power Reactors," Obninsk, Russia, October 16-19, 1994.
Feinroth 1993: Herbert Feinroth, AECL Technologies. Personal communication.
Feinroth 1994: Herbert Feinroth, AECL Technologies. Personal communication.
GA 1992: General Atomics. "Excess Plutonium Disposition: Modular High Temperature Gas-Cooled Reactor (MHTGR)." Draft Report, prepared for the Fission Working Group of the DOE Plutonium Dispositions Task Force, Rev. 4A, General Atomics, San Diego, Calif., December 21, 1992.
GA 1993a: General Atomics. "Utilizing the Modular Helium Reactor to Demilitarize Excess Weapons-Grade Plutonium." Briefing to Committee on International Security and Arms Control Panel on Reactor-Related Options, Washington, D.C., January 13, 1993.
GA 1993b: General Atomics. MHTGR Plutonium Consumption Study—Final Report. Contract DE-AC03-895F-17885. San Diego, Calif.: General Atomics, May 14, 1993.
GA 1994: General Atomics. MHTGR Plutonium Consumption Study, Phase II Final Report. GA/DOE-051-94. San Diego, Calif.: General Atomics, April 29, 1994.
Gacesa et al. 1983: M. Gacesa et al. "CANDU Design, Fuel Performance, and Operating Experience," AECL-MISC-250 (Rev. 1), 1. J. Hastings, Ed., 1983.
Garwin 1995: Richard L. Garwin. "Beyond the 'Spent Fuel Standard': Two Interpretations of 'Elimination' of Excess Weapons Plutonium." Unpublished paper available from the Committee on International Security and Arms Control, National Academy of Sciences, Washington, D.C., 1995.
Gat undated: Uri Gat. "Molten Salt Reactors for Burning Dismantled Weapons Fuel." Draft manuscript.
Gat 1986: Uri Gat. “The Ultimate Safe (U.S.) Reactor—A Concept for the Third Millennium." Paper presented at the 4th International Conference on Emerging Nuclear Safety Systems, Madrid, Spain, 1986.
Gat 1993: Letter from Uri Gat, Oak Ridge National Laboratory, to Duane Hanson, Idaho National Engineering Laboratory, March 1, 1993.
Gat et al. 1992: Uri Gat, J.R. Engel, and H.L. Dodds 1992. "Molten Salt Reactors for Burning Dismantled Weapons Fuel." Nuclear Technology 100(December), 1992.
GE 1994: Study of Plutonium Disposition Using Existing GE Advanced Boiling- Water Reactors. Contract DE-AC03-93SF19681. San Francisco, Calif: General Electric, June 1, 1994.
Glasstone 1955: Samuel Glasstone. Principles of Nuclear Reactor Engineering. New York: D. Van Nostrand Co., Inc., 1955.
Gleukler and Quinn 1994: Emil L. Gleukler and James E. Quinn, GE-Nuclear Energy. "The Advanced Liquid Metal Reactor Actinide Recycle System." Paper presented at the Ninth Pacific Basin Nuclear Conference, Sydney, Australia, May 1994.
Hebel et al. 1978: L.C. Hebel, E.L. Christensen, F.A. Donath, W.E. Falconer, L.J. Lidofsky, E.J. Moniz, T.H. Moss, R.L. Pigford, T.H. Pigford, G.I. Rochlin, R.H. Silsbee, and M.E. Wrenn. "Report to the American Physical Society by the Study Group on Nuclear Fuel Cycles and Waste Management." Review of Modern Physics 50(1), Part II, 1978.
Hibbs 1993: Mark Hibbs. "Waste Disposal Top Priority For Back End, Nikipelov Says." Nuclear Fuel, July 19, 1993.
IAEA 1992a: International Atomic Energy Agency. Ranking of Safety Issues for WWER-440 Model 230 Nuclear Power Plants. IAEA-TECDOC-640. Vienna: IAEA, 1992.
IAEA 1992b: International Atomic Energy Agency. The Safety of WWER-440 Model 230 Nuclear Power Plants. IAEA Special Report. Vienna: IAEA, 1992.
INEL 1993a: D.A. Brownson, D.J. Hanson, H.S. Blackman, R.L. Bratton, C.H. Cooper, K.S. Dawson, Jr., L.B. Lundberg, C.S. Olsen, A.M. Ougouag, P.D. Wheatley, R.E. Felt, and R.N. Henry. Weapons-Grade Plutonium Dispositioning, Vol. 2. Comparison of Plutonium Disposition Options. DOE/ID-10422. Idaho Falls, Idaho: Idaho National Engineering Laboratory, June 1993.
INEL 1993b: J.M. Ryskamp, B.G. Schnitzler, C.D. Fletcher, R.L. Moore, M.J. Gaeta, L.B. Lundberg, S.N. Jahshan, G.S. Chang, A.M. Ougouag, J.W. Sterbentz, C.A. Wemple, V.F. Dean, D.W. Akers, J.P. Sekot, and D.A. Brownson. Weapons-Grade Plutonium Dispositioning, Vol. 3: A New Reactor Concept Without Uranium or Thorium for Burning Weapons-Grade Plutonium. DOE/ID-10422. Idaho Falls, Idaho: Idaho National Engineering Laboratory, June 1993.
INEL 1993c: J.W. Sterbentz, C.S. Olsen, and U.P. Sinha. Weapons-Grade Plutonium Dispositioning, Vol. 4: Plutonium Dispositioning in Light-Water Reactors. DOE/ID-10422. Idaho Falls, Idaho: Idaho National Engineering Laboratory, June 1993.
JASON 1994: J. Cornwall, D. Eardley, R. Garwin, S. Koonin, G. MacDonald, R. Novick, W. Panofsky, S. Treiman, and F. Zachariasen. "Accelerator Based Conversion of Plutonium." JSR-94-310. JASON, MITRE Corporation, McLean, Virginia, November 22, 1994.
Koopman et al. 1992: Ronald P. Koopman, Donald L. Goldman, James A. Hassberger, Ernie Hill, Lawrence D. Ramspott, William G. Sutcliffe, Carl E. Walter and others from the Lawrence Livermore National Laboratory Fissile Materials Disposition Working Group, and A. David Rossin. Disposition of Plutonium and Highly Enriched Uranium from Nuclear Weapons, Part I—Issues and Options. Draft, Lawrence Livermore National Laboratory, Livermore, Calif., December 11, 1992.
Kudriavtsev 1993: E.G. Kudriavtsev. "Russian Prospects for Plutonium Accumulation and Utilization." Paper presented to International Atomic Energy Agency meeting on Problems of Separated Plutonium, April 1993.
Lane et al. 1958: James A. Lane, H.G. McPherson, and Frank Masian, eds. Fluid Flow Reactors. Reading, Mass.: Addison Wesley Pub. Co., Inc., 1958.
Levina et al. 1994: I.K. Levina, V.V. Saprykin, and A.G. Morozov. "The Safety Criteria and VVER Core Modification For Weapon Plutonium Utilization." Paper presented at NATO Advanced Research Workshop on "Mixed Oxide Fuel (MOX) Exploitation and Destruction in Power Reactors," Obninsk, Russia, October 16-19, 1994.
Lotts et al. 1992: A.L. Lotts, W.D. Bond, C.W. Forsberg, R.W. Glass, F.E. Harrington, G.E. Michaels, K.J. Notz, R.G. Wymer. Options for Treating High-Temperature Gas-Cooled Reactor Fuel for Repository Disposal. ORNL/TM-12027. Oak Ridge, Tenn.: Oak Ridge National Laboratory, February 1992.
Ludewig 1993a: Hans Ludewig, Brookhaven National Laboratory. Letter to Duane Hanson, Idaho National Engineering Laboratory, March 4, 1993.
Ludewig 1993b: Personal communication from Hans Ludewig, Brookhaven National Laboratory, April 7, 1993.
Ludewig 1993c: Personal communication from Hans Ludewig, Brookhaven National Laboratory, August 11, 1993.
Lyman 1994: Edwin S. Lyman. Interim Storage Matrices for Excess Plutonium: Approaching the "Spent Fuel Standard" Without the Use of Reactors. PU/CEES No. 286. Princeton, N.J.: Center for Energy and Environmental Studies, Princeton University, August 1994.
McNeese and Rosenthal 1974: L.E. McNeese and M.W. Rosenthal. "MSBR: A Review of its Status and Future." Nuclear News, September 1974.
McPherson 1985: H.G. McPherson. "The Molten Salt Reactor Adventure." Nuclear Science and Engineering 90:374-380, 1985.
Mikhailov et al. 1994: V.N. Mikhailov, V.V. Bogdan, V.M. Murogov, V.S. Kagramanian, N.S. Rabatnov, V. Ya. Rudneva, and M.F. Troyanov. "Utilization of Plutonium in Nuclear Power Industry of Russia." Paper presented to the "International Policy Forum: Management and Disposition of Nuclear Weapon Materials," Leesburg, Virginia, March 8-11, 1994.
Minkov et al. 1990: V.E. Minkov, G.A. Lambert and W.R. Simmons. Design Description of Soviet Liquid Metal Cooled Fast Breeder Reactors. ANLSETC 90-1, Argonne, III.: Argonne National Laboratory, February 1990.
Miraglia 1994: Frank Miraglia, Deputy Director of Nuclear Regulation, Nuclear Regulatory Commission. Personal communication, June 2, 1994.
Mitenkov et al. 1992: F.M. Mitenkov, O.B. Samoilov, Y.K. Panov, U.S. Cuul, L.N. Flerov, N.N. Ponomarev-Stepnoy, N.E. Kukharkin, and Y.G. Nikiporets. "VPBER-600 Enhanced Safety Water Cooled Medium Power Reactor." Paper presented at the International Conference on Design and Safety of Nuclear Power Plants, Tokyo, October 1992.
Miura et al. 1992: M. Miura, T. Inagaki, M. Hamada and Muramatsu, JAPC. "Present Status of DFBR Design in Japan." Paper presented at the International Conference on Design and Safety of Nuclear Power Plants, Tokyo, October 1992.
NAS 1994: National Academy of Sciences, Committee on International Security and Arms Control. Management and Disposition of Excess Weapons Plutonium. Washington, D.C.: National Academy Press, 1994.
National Research Council 1992: National Research Council, Committee on Future Nuclear Power Development. Nuclear Power: Technical and Institutional Options for the Future. Washington, D.C.: National Academy Press, 1992.
National Research Council (forthcoming): National Research Council, Committee on Separations Technology and Transmutation Systems. Report of the Committee on Separations Technology and Transmutation Systems. Washington, D.C.: National Academy Press, forthcoming.
NEI 1985: "The N4 Plant." Nuclear Engineering International, February 1985.
NEI 1988: "The British PWR." Supplement to Nuclear Engineering International, May 1988.
NEI 1993: World Nuclear Industry Handbook 1993. Nuclear Engineering International Special Publications, 1993.
Nigon and Golinelli 1994: Jean-Louis Nigon and Claude Golinelli. "MOX in France: Domestic Programme and MELOX Plant." Paper presented at NATO Advanced Research Workshop on "Mixed Oxide Fuel (MOX) Exploitation and Destruction in Power Reactors," Obninsk, Russia, October 16-19, 1994.
Novikov et al. 1994: A.N. Novikov, V.V. Saprykin, A.A. Suslov, and A.P. Lazarenko. "Use of MOX (R-Pu and W-Pu) Fuel in VVER-1000 (Neutron-Physical Aspects of Possibilities)." Paper presented at NATO Advanced Workshop on "Managing the Plutonium Surplus: Applications and Options," Royal Institute of International Affairs, London, January 24-25, 1994.
NRDC 1993: C.E. Paine and T.B. Cochran. "The Once-Through LWR/MOX Option for Disposing of Weapon-Grade Plutonium." Paper presented to the Committee on International Security and Arms Control, National Academy of Sciences, Washington, D.C., April 1993.
Nuclear News 1992a: "The Modular High-Temperature Gas-Cooled Reactor." Nuclear News, 78-79, December 1992.
Nuclear News 1992b: "The New Reactors." Nuclear News, September 1992.
Nuclear News 1993: "World List of Nuclear Power Plants." Nuclear News, September 1993 .
Nucleonics Week 1994: "Palo Verde Back at Full Power, but Six-Month Inspections Continue." Nucleonics Week, July 14, 1994.
Omberg and Walter 1993: Ronald P. Omberg and Carl E. Walter. Disposition of Plutonium from Dismantled Nuclear Weapons: Fission Options and Comparisons. UCRL-ID-113055. Livermore, Calif.: Lawrence Livermore National Laboratory, February 5, 1993.
OTA 1994: Office of Technology Assessment. Technical Options for the Advanced Liquid Metal Reactor—Background Paper. OTA-BP-ENV-126. Washington, D.C.: U.S. Government Printing Office, May 1994.
Redding and McGregor 1993: John Redding and Paul McGregor, General Electric. "Advanced Nuclear Plants: Meeting the Challenge." Paper presented at the American Nuclear Society winter meeting, San Francisco, November 1993.
Sako et al. 1992: K. Sako, K. Higashikubo, Y. Isozaki, K. Kaneko, K. Kato, Y. Kawauchi, J. Oda, H. Ogata, T. Oikawa, and M. Togasaki. "Passive Safe Reactor SPWR." Paper presented at the International Conference on "Design and Safety of Nuclear Power Plants," Tokyo, October 1992.
Sauerbrun 1993: T.J. Sauerbrun. "Cost and Schedule Estimates for Plutonium Burner Based on Comparative Cost Analysis Method." Idaho National Engineering Laboratory, April 21, 1993.
Schleicher et al. 1992: Robert W. Schleicher, David D. Kapich, and Klaus T. Etzel. "Modular HTGR Gas Turbine Power Plant." GA-A20890, March 1992. Paper presented at the 27th Intersociety Energy Conversion Engineering Conference, August 1992.
Schmeidel 1992: Peter Schmeidel, Siemens. Presentation to the International Workshop on the Disposal of Plutonium, Bonn, June 16, 1992 (cited in Berkhout et al. 1992).
Schmitt 1993a: C.H. Schmitt, Deputy Director, Office of Naval Reactors, U.S. Department of Energy, Personal communication, April 22, 1993.
Schmitt 1993b: C.H. Schmitt, Deputy Director, Office of Naval Reactors, U.S. Department of Energy, Personal communication, August 11, 1993.
Schmitt 1993c: C.H. Schmitt, Deputy Director, Office of Naval Reactors, U.S. Department of Energy, Personal communications, August 16, 18, and 19, 1993.
Schmitt 1994: C.H. Schmitt, Deputy Director, Office of Naval Reactors, U.S. Department of Energy. Letter to J.F. Ahearne, May 11, 1994.
Shapiro et al. 1977: N.L. Shapiro, R.R. Rec, and R.A. Matzie. Assessment of Thorium Fuel Cycles in Pressurized Water Reactors. EPRI NP-359. Palo Alto, Calif.: Electric Power Research Institute, 1977.
Thomas 1994: Wolfgang Thomas. "Use of MOX in Existing LWRs in Germany." Paper presented at NATO Advanced Research Workshop on "Mixed Oxide Fuel (MOX) Exploitation and Destruction in Power Reactors," Obninsk, Russia, October 16-19, 1994.
Till and Chang 1988: C.E. Till and Y.I. Chang. "The Integrated Fast Reactor." Advances in Nuclear Science & Technology 20:127-154, 1988.
Till and Chang 1990: C.E. Till and Y.I. Chang. "Actinide Recycle Potential in the Integral Fast Reactor (IFR) Fuel Cycle." Paper presented at the 1990 winter meeting of the American Nuclear Society, Washington, D.C., 1990.
Turk and Matzie 1992: Richard S. Turk and Regis A. Matzie, ABB-CE. "System 80+." Nuclear Engineering International, November 1992.
USDOE 1993: U.S. Department of Energy Plutonium Disposition Study, Technical Review Committee Report, Vol. I and II. Washington, D.C.: U.S. Department of Energy, July 2, 1993.
USNRC 1976: U.S. Nuclear Regulatory Commission. Final Generic Environmental Statement on the Use of Recycle Plutonium in Mixed Oxide Fuel in Light-Water Cooled Reactors. NUREG-0002. Springfield, Va.: National Technical Information Service, August 1976.
Weinberg 1970: A. Weinberg, ed. Nuclear Applications and Technology 8:102219, 1970.
Westinghouse 1994: Plutonium Disposition in Existing Pressurized Water Reactors. Contract No. DOE-ACO3-93SF19683. San Francisco, Calif.: Westinghouse Electric Corporation, June 1, 1994,
Wilcox 1994: Paul Wilcox. "Use of Plutonium in the UK." Paper presented at NATO Advanced Research Workshop on "Mixed Oxide Fuel (MOX) Exploitation and Destruction in Power Reactors," Obninsk, Russia, October 16-19, 1994.
Wymer et al. 1992: R. G. Wymer, H.D. Benglesdorf, G.R. Choppin, M.S. Coops, J. Guon, K.K.S. Pillay, and J.D. Williams. An Assessment of the Proliferation Potential and International Implications of the Integral Fast Breeder Reactor. Prepared by Martin Marietta International Technology
Programs Division for the U.S. Department of Energy and the U.S. Department of State. Oak Ridge, Tenn.: Oak Ridge National Laboratory, May 1992.
Yamano 1994: Tomohiro Yamano. "Japanese Nuclear Fuel Cycle Plutonium Utilization Policy." Paper presented at NATO Advanced Research Workshop on "Mixed Oxide Fuel (MOX) Exploitation and Destruction in Power Reactors," Obninsk, Russia, October 16-19, 1994.