U.S.-German Cooperation in the Elimination of Excess Weapons Plutonium (1995)

A vast majority of existing reactors in the world are light water reactors (LWRs) burning low enriched uranium (LEU) fuel. Nearly 80 percent of the 419 power reactors operating at the end of 1993 were LWRs and 8 percent were heavy water reactors (HWRs), generally burning natural uranium. Both of these reactor types operate with a thermal neutron spectrum and can also be fueled with mixed oxide (MOX) containing mixtures of uranium oxide and plutonium oxide. Most of these reactors operate on a “once-through” fuel cycle, that is, spent fuel is stored with the intent of eventual disposition in a geological repository. Alternatively, the spent fuel can be reprocessed, separating the reprocessed plutonium for use in MOX or possibly in breeder reactors once these have been successfully developed and built on a commercial scale. MOX can be fabricated using such reprocessed plutonium or using weapons-grade plutonium (WPu); the separated fission products (high-level radioactive waste, HLW) are then usually vitrified and stored for eventual emplacement in a mined geological repository.

There are significant physical differences between LEU fuel and MOX. These differences are important in that they determine how much plutonium can be loaded into MOX fuel elements and what fraction of the reactor core can use MOX fuel rods in place of LEU. These quantities in turn determine how many reactors are required for what period of time to dispose of a given quantity of excess WPu. To give guidance for such estimates, one should be reminded that roughly one ton of either plutonium or U²³⁵ is consumed to generate one gigawatt year of electricity.

The significant differences between plutonium and uranium-based reactor fuels are the following:

1. Delayed Neutrons. Delayed neutrons are important since their presence makes it possible to control the reactivity in a reactor through the mechanical insertion of control rods. About 0.7 percent of all the neutrons from the thermal fission of U²³⁵ are delayed, while only about 0.2 percent are delayed from Pu^239. Thus control rod performance must be enhanced if plutonium fuels are used.

2. Neutron Spectrum. Since plutonium has a higher absorption cross section for thermal neutrons compared to U²³⁵, the average energy of the neutron spectrum in reactors using plutonium fuels is higher. Thus the capture of neutrons by control rods is reduced if plutonium fuels are used, thus further reducing the effectiveness of the control rods.

3. Resonance. Pu²³⁹ has a peak in its absorption (called a resonance) at neutron energies near 0.3 eV. This has the result that under certain conditions the temperature coefficient of reactivity of the reactor can become positive, as the temperature increases more neutrons are being pushed into the resonance region, increasing reactivity still further.

The above differences have to be taken into account in reactor design initially intended for the use of MOX or for modifying existing reactors for the use of MOX. The effects noted above

can be managed in a number of ways, such as increasing the number of control rods or their speed of insertion, adding so-called burnable poisons, which are materials of high neutron absorption cross-section that absorb the extra neutrons, in particular those in the resonance region, and other methods.

The plutonium content in spent fuel will generally be larger if MOX is used in place of LEU. The reason is that the unused plutonium from the MOX will add to the plutonium that is bred from the U²³⁸, which is contained both in LEU and MOX fuel. However the plutonium in spent fuel will of course be reactor-grade (RPu), that is, its isotopic composition will be changed, and the spent fuel will meet the “spent fuel standard” in that the RPu is no more accessible to potential diversion than the plutonium from thermal reactors burning LEU or natural uranium. It may appear that the potentially larger plutonium content of spent fuel is a negative factor arguing against the use of MOX for disposing of WPu. This is clearly not the case for two reasons: First the amount of RPu in spent fuel, if MOX made from WPu is used, is under all circumstances a tiny fraction of the RPu now contained in the world ’s inventories from commercial nuclear fuel. Second, the total amount of plutonium contained in spent fuel, if MOX fabricated from WPu is used, is less, relative to the total amount of the original WPu put into the fuel, plus the plutonium produced in the spent fuel had the same amount of electric power been generated from LEU. In short, the plutonium content of the spent fuel is not a useful discriminant among alternate disposition approaches.

Some currently operating reactors are already designed for using MOX for all their fuel elements and others not now burning MOX have been demonstrated to be usable for full MOX cores. In the former category is the American pressurized water reactor (PWR) designed by Combustion Engineering, called the System-80. It incorporates the additional control rods and increased neutron absorber in the coolant required to permit full MOX operation. Also the Canadian deuterium-uranium (CANDU) reactors have been operated on experimental basis with MOX and the contractors operating CANDU conclude that the design safety margins in CANDU using full MOX loading should not be significantly different from the margin in the current system fueled by natural uranium.

Actual commercial experience with MOX operations exists only in Europe since reprocessing of spent fuel is not practiced (licensed) in the United States or Canada and since use of MOX based on surplus WPu remains a matter for the future. In Germany, the first experience with the use of MOX was in boiling water reactors (BWRs) in 1966 with the experimental reactor Kahl (VAK). Commercial operation with PWRs with MOX started in the reactor Obrigheim (KWO), which operated until 1980. Since 1981 MOX fuel has been used in additional PWRs. In total about 67.5 tons of plutonium has been processed in more than 100,000 fuel rods containing MOX. Extensive experience with a large variety of parameters, such as fuel composition and varying degrees of burn-up, has accumulated. These German reactors have been licensed for MOX fractions up to 50 percent. Table B-1 gives an overview of these reactors. The actual percentage of MOX use has been considerably less than that shown in the table. The reason is that not enough fuel was available. In principle, it could also be possible to license the BWRs for

partial-core MOX use. This has not happened in the past because of the limited MOX production capacities. Table B-2 gives an overview of the BWRs.

As of the end of 1993, eight LWRs in France and two in Switzerland were using MOX fuel and more are licensed to do so in Western Europe. Belgium and Japan are planning to begin loading MOX fuel in commercial reactors later in this decade.

In summary, the use of MOX in LWRs in Europe, using partial MOX loads in reactor cores, has become regular commercial practice. Details of this favorable experience were presented in considerable detail at the GAAC workshop. In the United States and Canada, no reactors are licensed for MOX operation but the licensing process had been pursued in the past. If it were decided to use MOX in American or Canadian reactors, the licensing process could be relatively expeditious.

While actual operating experience with MOX had been restricted to partial MOX loads and while most existing LWRs were originally perceived to permit only partial loadings, recent analyses by the major U.S. reactor vendors (General Electric and Westinghouse) have indicated that, with relatively minor modifications and careful management of the location of MOX fuel bundles that have experienced varying degrees of burn-up, almost all existing LWRs could use a full MOX complement. These findings are only preliminary and would have to be put on a firm technical basis for an actual licensing procedure.

In view of the above, it is difficult to forecast precisely how many reactor years of MOX use would be required to dispose of the nominal 100 tons of surplus Russian WPu or the corresponding amount of U.S. excess inventory. If full MOX loading were possible, then four to six reactors could handle 100 tons in about 20 years. Partial loadings would increase the number of required reactors or the time needed for a campaign accordingly. Extensive calculations on specific combinations of actual reactors, assuming various percentages of plutonium in MOX, and assuming various fractions of MOX fuel loading in the reactors, are available in the literature.¹

Estimates of the costs of using WPu in existing LWRs or CANDUs are given in the report of the NAS Panel on Reactor-Related Options for the Disposition of Excess Weapons Plutonium.² Estimates have been somewhat contentious due to the variety of the assumptions made and because of the variability of cost experience. There is general agreement that the cost of fueling a reactor with MOX fabricated from free WPu today would be somewhat higher than fueling the same reactor with fully paid-for LEU; fully-paid for means including the costs of mining, processing, enriching, and fabricating the LEU fuel. Thus the use of WPu as a substitute for LEU in existing reactors is likely to require a subsidy despite the inherent fuel value of WPu. The cost impact of such a differential on the total cost of electricity is minimal, however.

Fabricating MOX fuel is considerably more expensive than fabricating LEU of comparable energy value. The reason is that MOX fabrication, due to the radioactivity, toxicity, and safeguarding requirements, has to be much more automated than that for LEU and requires additional shielding and other facilities and highly secure procedures. For these reasons LEU fuel fabrication facilities cannot readily be modified to MOX fabrication.

There are no operating MOX fabrication facilities in the United States. A partially completed MOX fabrication facility exists at the U.S. DOE Hanford site known as the Fuel and Materials Examination Facility (FMEF). The FMEF was built in the late 1970s and early 1980s. This facility could be completed to produce about 50 tons of MOX fuel per year, containing roughly 3 tons of WPu.

There are extensive operating MOX facilities in Europe. While not now using MOX fuel in its own commercial reactors, Belgium today provides a significant fraction of the world’s currently operating MOX fabrication capability. Its MOX fabrication services are marketed in conjunction with France. In Germany a modest-scale MOX fabrication plant at Hanau operated for several years before losing its license and a larger plant is nearly completed. This small plant processed about 8.5 tons of RPu into MOX fuel before shutdown. The larger plant has a capacity to process about 5 tons of WPu per year and has now been licensed for RPu after considerable litigation and political controversy. Operation of that plant has still not been possible, however, since the authorities in the State of Hesse are opposed to it, and have used their implementation authority to slow progress. The total investment in this plant is quite large (approaching $1 billion), and to maintain it in a standby basis requires approximately DM 10 million per month. Thus a final decision on the future of this plant is imminent.

Britain, while not having a domestic MOX use program, is building a large MOX fabrication plant for foreign customers that will be completed soon. Japan is constructing a substantial MOX fabrication facility for future use. France is also building a substantial MOX fabrication facility. Russia has two partially completed MOX fabrication facilities, but these are suitable only for fabricating MOX for fast reactors and do not meet internationally accepted safety criteria. Accordingly, it is doubtful that they will be completed. The Russian-German feasibility study recommended abandoning them.

With exception of the almost-completed Hanau plant, most of the existing and projected MOX fabrication capacity in Europe is fully contracted to fabricate reprocessed commercial reactor fuel. Therefore, in the absence of new facilities being constructed, substantial quantities of WPu withdrawn from excess nuclear weapons could be fabricated under MOX fuel only if the U.S. facility (FMEF) were completed, at an estimated cost of about $100 million, or if the obstacles now preventing the Hanau facility from becoming operational could be removed.

An alternative approach of substituting WPu for the planned MOX fabrication of already contracted RPu reprocessed from commercial spent fuel has been suggested, but it appears unrealistic in view of the existing commitments. Moreover, unless the commercial spent fuel were not reprocessed, it would simply substitute WPu for the accumulating inventories of reprocessed RPu from commercial spent fuel; since the latter also constitutes a proliferation risk, such a substitution does not seem to be to offer any nonproliferation advantage.

It should be noted that the controversial issue concerning reprocessing of commercial spent fuel is not involved in the question of whether, where, and how WPu from excess nuclear weapons should be fabricated into MOX. The existence of the WPu is a physical reality and it has already been reprocessed. Thus, the issue of MOX fuel fabrication deals with the question of whether MOX fabrication facilities should be made available for purposes of disarmament, and the question of fabricating MOX in the civilian nuclear power fuel cycle is an independent issue.