The U.S. Department of Energy announced on January 14, 1997 that it will study a "dual-track" approach to put approximately 50 metric tons of plutonium rendered surplus by the end of the Cold War into forms not usable for making nuclear weapons. ² One "track" would vitrify plutonium -- that is, mix it with molten glass and other materials. The DOE proposes to use this for 8 to 17 metric tons of surplus weapons plutonium. The other track would convert plutonium into a fuel for nuclear reactors. This involves putting plutonium into an oxide chemical form, mixing it with uranium oxide, and fabricating it into ceramic fuel pellets (called MOX fuel for short). MOX fuel would be put into reactor fuel rods and loaded into reactors as a complete or partial substitute for the uranium fuel currently used.

While much of the official discussion about MOX is that it would "burn" the plutonium, in reality plutonium is both consumed ("burned") and produced in nuclear reactors. ³ MOX irradiated (or "spent") reactor fuel would still contain from 40 percent to over 70 percent of the original amount of plutonium after it is discharged from the reactor (see table in MOX spent fuel section). ⁴ This spent fuel contains highly radioactive materials resulting from fission and other nuclear reactions during reactor operation. The main function of both the vitrification and MOX options is not to get rid of all the plutonium. Neither method does that. Rather it is to:

• mix plutonium with other materials so that it would be very difficult to re-extract for use in weapons;
• prevent diversion of plutonium by putting it into highly radioactive storage forms that would be lethal to anyone wanting to steal it. This is automatically accomplished in the case of MOX spent fuel which is mixed with fission products. Plutonium can also be mixed with fission products during vitrification.

This article discusses technical issues related to the use of MOX fuel derived from weapons plutonium in nuclear power reactors. ⁵ Some economic issues are also discussed.

MOX fabrication

MOX fuel has never been fabricated on an industrial scale from weapons-grade plutonium. Current industrial MOX facilities use plutonium dioxide derived from facilities that reprocess spent power reactor fuel (called reactor-grade plutonium). There are some important differences.

Commercial reprocessing plants currently use aqueous technology (that is, acids and other liquid solvents) to separate plutonium and uranium in spent fuel from fission products and from each other (see SDA Vol. 5 No. 1). The final product is a plutonium dioxide power that can be directly used in MOX fuel production. ⁶In contrast, most military plutonium is in the form of "pits" which consist of plutonium metal with small quantities of other materials. In the United States (and elsewhere) weapons plutonium is alloyed with up to one percent gallium. Since relatively pure plutonium dioxide powder is needed for MOX fuel fabrication, the weapons plutonium metal must both be purified and converted into oxide form (not necessarily in that order) before it can be used. It is particularly important to remove the gallium almost completely. Thus, MOX fuel fabrication from weapons-grade plutonium involves steps and processes that are not needed for reprocessed plutonium from power reactor fuel. (See article on gallium.)

The current processes for making weapons plutonium into suitable feed for a MOX fuel fabrication plant use aqueous technology similar to reprocessing. That is, they involve dissolution of plutonium pits in acid followed by purification of the plutonium and conversion into an oxide form. These aqueous processes produce large amounts of liquid radioactive wastes. For instance, one aqueous process would, for every 30 metric tons of weapons plutonium converted into plutonium dioxide, produce between 800,000 and 900,000 gallons of liquid wastes with specific radioactivity of 20 to 30 picocuries per liter. ⁷

Dry processes that could be used to make plutonium oxide and remove gallium have not yet been developed beyond the laboratory scale. They will take four to five years more to reach the industrial scale needed for plutonium disposition using MOX.

MOX has been made in the US only in small-scale glove-box facilities. In order to use the MOX option, the United States would have to construct a new fuel fabrication plant or complete the partially-finished Fuel Materials Examination Facility at the Hanford site in Washington state, built in the 1970s to produce breeder reactor fuel. Besides Hanford, a MOX plant could be built at the Pantex Plant in Texas, the Idaho National Engineering Laboratory, or the Savannah River Site in South Carolina. Facilities in Europe may be used for initial MOX fuel loadings, but a US MOX facility would eventually be built.

MOX Utilization

Eighteen power reactors in Germany, France, and Belgium are using MOX fuel. France plans to expand the number of reactors using MOX from nine to 16 reactors by the year 2000. All of these are light water reactors (LWRs). These reactors use ordinary water for slowing down the neutrons needed to maintain the nuclear chain reaction and for cooling the reactor.

In the United States, MOX fuel was used in tests in LWRs during the 1960s and 1970s. But drawing on the European experience, dozens of LWRs could potentially be used for plutonium disposition (see below for discussion of safety issues). The time it would take to convert plutonium into non-weapons-usable irradiated fuel in reactors depends on a number of factors:

• the number, size, and type of reactors used, and average reactor power output
• the percentage of plutonium in the MOX fuel
• the percentage of the reactor core that is loaded with MOX fuel

With one-third MOX cores, and 2.5 percent plutonium in the MOX, it would take 8 reactors (of 1,000 megawatts electrical each) about 30 years to complete disposition of 50 metric tons of plutonium. The number of years would be reduced proportionally to the increase in MOX core loading, the number of reactors used, and their power output. Thus, three reactors operating on a full MOX core with 6.8 percent plutonium could complete the disposition in about 10 years.

MOX fuel has not been used in Russian LWRs at all, but has been tested in other reactor designs, including the breeder (or fast neutron) reactor. In 1995, the US National Academy of Sciences (NAS) panel on reactor options for plutonium disposition determined that, for safety reasons, the VVER-440 reactors (smaller Russian light water reactors) and the carbon-moderated RBMK reactors (of the Chernobyl type) were unsuitable for MOX fuel use. Further, while the larger light water reactors, known as VVER-1000 reactors, could be considered for MOX use because their safety standards are higher than other Russian reactors, these reactors "do not currently meet international safety standards," according to the same NAS study. ⁸ However, the NAS notes that these reactors are being upgraded with international assistance. There are seven reactors of this type in Russia and ten in Ukraine.

The Russian Ministry of Atomic Energy, Minatom, had not seriously considered the use of MOX in LWRs until the US plutonium disposition program created greater incentives to look at this option. Minatom generally favors fast breeder reactors for the use of plutonium fuel. This preference arises from the fact that Russia considers plutonium an energy treasure and still has the long-term goal of creating a plutonium-fueled nuclear energy system using fast breeder reactors and reprocessing plants. Russia is also considering pursuing plutonium disposition using its one fast breeder reactor, BN-600, though that reactor now uses uranium fuel. Further breeder reactor construction in Russia is stalled due to lack of funds.

Canadian heavy water reactors (called "CANDU" reactors, which use natural uranium as fuel and heavy water as a moderator and coolant) are also being considered for disposition of US surplus military plutonium and also possibly for the Russian surplus. Unlike LWRs, which are shut down periodically for refueling, these reactors are continually fueled.

CANDU reactors would use 100 percent MOX cores. Atomic Energy of Canada Limited (AECL), which is the vendor of Canadian reactors, reported to the US NAS committee on plutonium disposition that it has extensive experience in testing the use of MOX fuel containing from 0.5 to 3 percent plutonium. According to the AECL, CANDU reactors can use 100 percent MOX cores without physical modification, ⁹ but new licensing would be required because no CANDU reactors are currently licensed to use MOX fuel. CANDU reactors could accommodate 100 percent MOX cores because they have adequate space for any additional control blades (similar to control rods) that may be needed.

CANDU reactors appear to have a number of significant advantages in the use of MOX fuel in terms of controllability. The power production per unit of fuel would be higher with MOX fuel than with natural uranium fuel. With higher power production, the volume of high-level radioactive waste produced by these reactors would be smaller than that now produced by CANDU reactors. Yet CANDU reactors also possess many disadvantages, such as the need for international transport of MOX fuel, which can be chemically separated into uranium and weapons-usable plutonium in a relatively straightforward manner. Use of CANDU reactors may also require production of a greater volume of MOX fuel than use of LWRs, since the fuel would contain between 1.5 percent and 2.7 percent plutonium, ¹⁰ rather than the 4 percent or more possible in light water reactors. Canada would use MOX made in the United States.

Specific Plans for MOX in the United States

The MOX options that DOE is considering for disposition of surplus weapons plutonium are:

• existing light water reactors (LWRs) in the United States
• partially-completed LWRs which would be completed for the purpose of plutonium disposition
• evolutionary LWRs (new reactors built by the DOE for the explicit purpose of plutonium disposition)
• CANDU reactors

Eighteen US utilities have expressed interest in using MOX fuel. Some of them have also indicated an interest in making tritium for the nuclear weapons program. (An interactive map of these utilities and reactors is given here.)

Light Water Reactor Safety and Licensing Issues related to MOX

The vast majority of LWRs were not designed to use plutonium as a fuel. While both plutonium-239 and uranium-235 are fissile materials that generate similar amounts of energy per unit weight, there are a number of differences between them as reactor fuels that affect reactor safety. The basic set of concerns relates to control of the reactor. The chain reaction in a reactor must be maintained with a great deal of precision. This control is achieved using control rods usually made of boron and (in pressurized water reactors) by adding boron to the water. Control rods allow for increases and decreases in the levels of reactor power and for orderly reactor shut-down. They prevent runaway nuclear reactions that would result in catastrophic accidents.

It should be noted that while all commercial LWRs have some amount of plutonium in them which is made during the course of reactor operation from uranium-238 in the fuel, the total amount of plutonium is about one percent or less when low enriched uranium fuel is used. When MOX fuel is used, the total amount of plutonium would at all times be considerably higher. It is this difference that creates most reactor control issues.

Changing the fuel can affect the ability of the control rods to provide the needed amount of reactor control. Hence, modifications to the reactor may be required before the new fuel can be used. Therefore, changing the fuel in any significant way also requires re-licensing of the reactor.

Several differences between the use of MOX fuel and uranium fuel affect safety:

• The rate of fission of plutonium tends to increase with temperature. This can adversely affect reactor control and require compensating measures (see box on reactor control). This problem is greater with MOX made with weapons-grade plutonium than that made with reactor-grade plutonium.
• Reactor control depends on the small fraction of neutrons (called delayed neutrons) emitted seconds to minutes after fission of uranium or plutonium. Uranium-235 fission yields about 0.65 percent delayed neutrons, but plutonium yields only about 0.2 percent delayed neutrons. This means that provisions must be made for increased control if plutonium fuel is used, if present control levels and speeds are deemed inadequate. (See box on reactor control.)
• Neutrons in reactors using plutonium fuel have a higher average energy than those in reactors using uranium fuel. This increases radiation damage to reactor parts.
• Plutonium captures neutrons with a higher probability than uranium. As a result, a greater amount of neutron absorbers are required to control the reactor.
• The higher proportion of plutonium in the fuel would increase the release of plutonium and other transuranic elements to the environment in case of a severe accident.
• Irradiated MOX fuel is thermally hotter than uranium fuel because larger quantities of transuranic elements are produced during reactor operation when MOX fuel is used.

Overall, the issues related to reactor control, both during normal operations and emergencies, are the most crucial. Most independent authorities have suggested that only about one third of the fuel in an LWR can be MOX, unless the reactor is specifically designed to use MOX fuel. However, there are some operational problems associated with using partial-MOX cores since MOX fuel is interspersed with uranium fuel. Their differing characteristics regarding control, radiation and thermal energy mean that there are non-uniform conditions in the reactor that can render operation and control more complicated. Some reactor operators claim they can use 100 percent MOX cores without needing to make physical changes to the reactor or control rods. The safety implications of such claims need to be independently verified.

Some newer reactors, however, have been designed for the use of a 100% MOX core because during reactor design appropriate provisions were made for additional control. There are only three reactors of this type in the US: the three System-80 reactors of the Arizona Public Services Company located at Palo Verde. These reactors are under consideration for disposition of surplus US plutonium. However, even if 100 percent MOX cores were allowed, the percentage of plutonium in the MOX would likely be on the low side, so that a larger amount of MOX fuel would have to be fabricated. Hence the advantages from the point of view of speed of disposition of such an approach may be relatively small.

MOX Spent Fuel

Using reactors to dispose of surplus weapons plutonium will not result in complete elimination of the plutonium. MOX spent fuel contains more plutonium and is thermally hotter than conventional spent fuel (that is, spent fuel resulting from loading an LWR with low enriched uranium fuel). Conventional spent fuel from light water reactors typically contains about one percent plutonium when it is withdrawn from the reactor. The amount of residual plutonium in MOX spent fuel would depend on the initial plutonium loading (percent of plutonium in the fuel), the burn-up of the fuel, and the configuration in which the fuel is used.

For light water reactors using MOX fuel, the NAS calculates that residual plutonium in the spent fuel would range from 1.6 percent (for a 33% MOX core with 4% plutonium loading) to 4.9 percent (for a 100% MOX core with 6.8% plutonium loading). Ranges of 2.5 percent to 6.8 percent plutonium loading have been suggested. ¹¹ In the case of a CANDU reactor using a 100% MOX core, the percentage of plutonium in MOX spent fuel would be between 0.8 and 1.4 percent for MOX fuel containing 1.2 percent and 2.1 percent plutonium, respectively. ¹² (See table.)

COMPARISON OF PLUTONIUM IN SPENT FUEL FOR VARIOUS REACTORS USING URANIUM AND MOX FUEL
TYPE OF FUEL	TYPE OF REACTOR	% OF FUEL THAT IS MOX	% PU IN FRESH FUEL	% PU IN SPENT FUEL
Uranium	LWR	NA	NA	0.9
MOX	LWR	33	4	1.6
MOX	LWR	100	4	2.6
MOX	ELWR	100	6.8	4.9
MOX	CANDU	100	1.2	0.8
MOX	CANDU	100	2.1	1.4
Adapted from: NAS 1995 p. 252 table 6.1.
LWR: light water reactor ELWR: evolutionary light water reactor CANDU: Canadian deuterium-uranium reactor (which would use MOX fuel) MOX: mixed [uranium-plutonium] oxide fuel

Technical issues related to MOX spent fuel disposal

Repository disposal of MOX spent fuel is complicated not only by the higher plutonium content in MOX, but by the larger quantities of transuranic elements in the spent fuel as well. This results in MOX spent fuel being thermally hotter than conventional spent fuel. The presence of greater amounts of transuranic radionuclides like americium-241 also cause persistent higher spent fuel temperatures, and cause the decay of thermal power level to be slower. MOX spent fuel use may therefore require that a host of issues be revisited, such as design of transportation and disposal canisters, and design of on-site spent fuel storage casks. For instance, the higher temperatures may cause storage problems at reactors that have limited storage room in their spent fuel pools. The higher temperature may also result in a need for more repository space, unless a repository is designed to take hotter fuel and withstand higher temperatures (a possibility being considered for Yucca Mountain). Greater repository space would result in proportionally higher repository disposal costs. In addition, if the amount of residual gallium in MOX spent fuel is too high, it may result in deterioration of the spent fuel cladding (see gallium article), create new issues in evaluating the suitability of a repository, and pose greater risk of groundwater contamination. There are some uncertainties as to the concentration of gallium that might adversely affect spent fuel integrity.

Financial Issues ¹³

Even though plutonium will be used to generate electricity in nuclear reactors, the use of MOX fuel will involve net costs. This is because it is more expensive to fabricate MOX fuel even when the plutonium is free than it is to purchase low-enriched uranium fuel, taking all costs, including raw material costs, into account. The cost of LEU fuel estimated by the NAS is about $1,400 per kilogram. MOX fuel fabrication using a new MOX plant was estimated at about $1,900. (Estimates in 1992 dollars). If the MOX fuel contains 5% plutonium, 50 metric tons of surplus weapons plutonium would yield 1,000 metric tons of MOX fuel. This would mean a fuel fabrication cost of about $1.9 billion in 1992 dollars, or over $2 billion in 1996 dollars.

The costs of MOX fuel fabrication may turn out to be higher than those estimated by the NAS. While an estimate of the costs of converting plutonium pits to oxide was included in the 1995 NAS report, there was no explicit treatment of the gallium problem. This was in part because DOE experts felt at the time that gallium might be left in the final MOX fuel. Subsequently, the severity of the problems created by gallium in the sintering process (the final step in MOX fuel fabrication) was discovered. Other potential problems that gallium could cause also came to light. Hence, it became necessary to separate the gallium from the plutonium, but these costs are not explicitly accounted for. In sum, the financial allowance for pit conversion to oxide made by the NAS may or may not be sufficient.

DOE policy appears to be that utilities would be sold MOX fuel at the cost of equivalent uranium fuel. Using the NAS cost estimates, this would make the net cost of making MOX about $500 million (for 50 metric tons of plutonium). DOE estimates of net MOX costs are generally lower. ¹⁴ In addition, there would be licensing costs for reactors, transportation and safeguard costs, and reactor modification costs (if such modifications are required).

It is difficult to estimate the total costs of plutonium disposition using MOX, but the DOE puts the estimate at about $2 billion for disposition in existing LWRs on the assumption that no subsidies to the utilities would be required. ¹⁵ However, the utilities want subsidies -- that is, they want compensation well beyond out-of-pocket costs. For instance, Jack Bailey, Vice-president of the Palo Verde nuclear plants, which can use 100 percent MOX cores (and are therefore leading candidates for MOX fuel use), stated his company's requirements for added compensation quite bluntly and publicly in March 1996:

According to a survey by General Electric, other utilities have also expressed requirements for compensation far in excess of direct cost reimbursement. Specifically, since many nuclear reactors will rapidly become uneconomical as electricity is deregulated in the next few years, they would require subsidies in order to be kept in operation for the MOX disposition track. ¹⁷ For licensing and/or safety reasons, the newer reactors are generally more likely to be selected for MOX use. But the newer reactors are far more expensive than the older ones, which would mean the MOX options could involve huge subsidies. Licensing delays would add to these costs.

Finally, the overall costs of MOX spent fuel disposal may be higher than that of uranium spent fuel, possibly by as much as a factor of two. ¹⁸ The Final Programmatic Environmental Impact Statement says nothing about added MOX spent fuel disposal costs. In fact, it assumes that there would be no added costs by stating that MOX spent fuel disposal costs are covered under the Nuclear Waste Policy Act, which did not anticipate higher costs associated with MOX spent fuel. However, as DOE's proposal now stands, any additional costs would be borne by the taxpayer.

Overall, the DOE estimates that using MOX fuel in existing reactors to dispose of 50 metric tons of weapons plutonium would cost about $2 billion, while vitrifying the plutonium would cost about $1.8 billion. ¹⁹ Given the many uncertainties surrounding plutonium disposition, this difference in cost estimates is not significant. However, as we have discussed above, DOE's estimates of MOX disposal costs will, in all likelihood, turn out to be severe underestimates. Therefore, the implication in DOE's analysis that MOX and vitrification disposition costs are comparable is likely to be wrong, and the MOX option will probably wind up being far more expensive.