IPFM International Panel on Fissile Materials

Limiting National Fissile-Material Production Capabilities

The crisis over Iran's uranium enrichment program and the controversy over Japan's new commercial reprocessing plant have each underscored the fact that the Nuclear Nonproliferation Treaty does not prevent non-nuclear weapon state parties to the treaty from deploying national uranium enrichment and reprocessing facilities or building up a huge stockpile of separated plutonium. NPT Article IV.1 states that:

"Nothing in this Treaty shall be interpreted as affecting the inalienable right of all the Parties to the Treaty to develop research, production and use of nuclear energy for peaceful purposes without discrimination and in conformity with Articles I and II of this Treaty."

As weapon states reduce their stockpiles, similar concerns could arise about the military potential of their enrichment and reprocessing facilities.

The debate therefore revolves around a Party's intentions, i.e. whether it intends to conform permanently with its nonproliferation commitments under Articles I and II. In the case of stockpiles of separated plutonium, it also revolves around concerns about the possibility of theft, which is also a concern in the weapon states.

As shown in Figure 1, eleven countries today have civilian enrichment and/or reprocessing plants in operation or under construction. Six of the eleven countries are nuclear-weapon states. The non-weapon states that have operating facilities are Brazil, Germany, Japan and the Netherlands. Iran is doing research and development with a small centrifuge cascade at Natanz and is planning to build and operate a much larger facility there.

Five non-weapon states that had operational pilot-scale facilities: Argentina and South Africa (enrichment), and Belgium, Germany and Italy (reprocessing) have suspended or terminated their programs.

Germany and the Netherlands are members of a multinational enrichment consortium, Urenco. This leaves Brazil, Iran and Japan as the only non-nuclear-weapon states with purely national fuel-cycle facilities.

There are a number of critical differences between uranium enrichment and reprocessing. One is that reprocessing is not currently an essential part of the fuel cycle of the for light-water reactors that now dominate civilian nuclear power. It can be postponed indefinitely by storing the spent fuel.

Uranium enrichment, however, is needed to supply the fuel for the reactors. The need for security of supply therefore provides a plausible reason for countries to build national enrichment plants. This was the justification offered by Japan, Brazil and Iran. France and the United States too are insisting on building new domestic enrichment plants designed by Urenco and using Urenco centrifuges, even though they could equally well invest in expanding the capacity of Urenco's existing enrichment plants.

FIGURE 1. Civilian enrichment and reprocessing plants worldwide. At present, Brazil, Iran, and Japan are the only non-weapons states with purely national fuel-cycle facilities. Germany and the Netherlands have enrichment plants but these are part of a multinational consortium.
Click on figure to go to IPFM Fissile Material Atlas and view interactive map.

Proliferation Dangers Associated with Centrifuge Enrichment

Centrifuges currently have decisively superior economics to other uranium enrichment technologies. They account for half of the world's enrichment capacity, and will account for all of it after France and the United States complete their current programs to replace their gaseous diffusion plants with centrifuge plants. Therefore, there is every reason for a country wishing to acquire an enrichment plant, to choose centrifuge technology.

Gas-centrifuge enrichment technology creates special proliferation concerns, however. First, because of its small inventory of uranium-hexafluoride, a centrifuge plant can convert rapidly from producing low-enriched uranium for power-reactor fuel to producing highly enriched uranium for weapons. Second, if a country wished to build a small clandestine centrifuge plant, it would be difficult to detect. A centrifuge plant uses relatively little power and leaks almost no gas to the atmosphere. This contrasts dramatically with the first uranium enrichment plants in the declared nuclear weapon states, which were gas-diffusion plants with huge inventories and power requirements.

Figure 2 shows France's Eurodif gas-diffusion plant in the back right (capacity: 8.5 million SWUs/yr). In the foreground are four full-sized 915 MWe nuclear power reactors, more than half of whose combined output is required to power the enrichment plant when it is operating at full capacity. The energy intensity of the plant is also dramatized by the enormous cooling towers required to remove the heat generated by the pumps that force uranium hexafluoride gas through thousands of diffusion barriers.

For contrast, Figure 3 gives a view of Urenco's centrifuge enrichment plant in the Netherlands (3.5 million SWUs/yr). The capacity is about half as large as that of the Eurodif plant but, because centrifuge enrichment requires only a few percent as much energy per separative-work unit (SWU), it requires neither a nearby power plant nor cooling towers to remove waste heat from the plant. From the air or space, the plant is not obviously distinguishable from any other factory.

FIGURE 2. France's Eurodif gas-diffusion uranium enrichment plant (large-area buildings in back) requires so much electrical power that it is co-located with a four-unit nuclear power plant.

FIGURE 3. Urenco's Almelo centrifuge enrichment plant has no associated power plants or cooling towers.

For a small enrichment plant, the situation is much much worse. It only requires an enrichment capacity of about 5000 SWUs/year, about 0.15 percent of the capacity of the Almelo plant, to produce enough weapon-grade uranium annually to make 25 kilograms of weapon-grade uranium -- enough for an implosion bomb. A gas-centrifuge plant of this size could be hidden relatively easily in a small, anonymous building or even underground. The floor area required could be contained in a square approximately 25 meters on a side. Such a plant would consume only about 100 kilowatts of electrical power.

Thus, once a country mastered the technology, it could, in principle, build a clandestine centrifuge-enrichment facility -- one of the possibilities driving concerns about Iran's centrifuge program.

Economic competitiveness, however, is a moving target. As a cumulative result of Urenco's long-term research and development program, each generation of its machines has had dramatically improved capacity and performance (see Figure 4). Urenco designs have made all other designs noncompetitive except for those fabricated in Russia, which adopted a different approach based on stacks of ever faster-spinning short centrifuges, while Urenco built each generation of centrifuges taller as well as faster.

There is therefore an economic incentive for even advanced countries to acquire centrifuge plants from Urenco or Russia. This is why both France and the United States are acquiring Urenco centrifuge plants. China has similarly built two centrifuge-enrichment plants using centrifuges supplied by Russia. The U.S. Enrichment Corporation hopes to leapfrog this competition by building a plant based on huge and costly centrifuges with enrichment capacities of 300-400 SWU/yr using technology developed by a $3 billion U.S. Department of Energy program. There is some skepticism, however, about its prospects for success.

FIGURE 4. The capacity of modern centrifuges is increasing. As shown here, Urenco has been able to improve the performance and increase the capacity of its machines dramatically. Such advanced machines could not be produced independently without a similarly dedicated long-term research and development effort. Building and successfully operating a first-generation machine, however, has become easier due to availability of high-precision tools and equipment.

Reprocessing

Whereas enrichment is essential to supply fuel for light-water reactors, reprocessing of spent fuel can be postponed indefinitely. Indeed, it is generally accepted today that, for the foreseeable future, reprocessing and plutonium recycle will be less economic than purchasing fresh low-enriched uranium fuel and storing the spent fuel. This is true, even in France, the country that is generally viewed as having the most successful reprocessing and plutonium recycle program. In 2000, the French government concluded that, even with its reprocessing and MOX-fuel fabrication plants paid for, France would save $4 to $5 billion over the remaining lifetime of its current fleet of power reactors if it stopped reprocessing in 2010.

Originally, interest in civilian reprocessing stemmed from programs in the industrialized countries to commercialize plutonium-breeder reactors. These reactors were to be fueled by the plutonium produced by neutron capture on the uranium isotope uranium-238. U-238 is 140 times more abundant than U-235, the primary fuel of current-generation reactors.

To provide plutonium for the initial cores of their planned breeder reactors, the major industrialized countries launched programs to harvest the plutonium contained in the light-water reactor spent fuel. The plutonium makes up about 1 percent of the heavy metal in the spent fuel. Britain and France used the expertise that they had developed in their weapons programs to build large-scale commercial reprocessing plants paid for by pre-paid reprocessing contracts from foreign utilities.

The breeder dream soon collapsed, however. The United States and Germany abandoned demonstration breeder reactor projects before they were completed. France, Japan and the United Kingdom completed demonstration reactors, but they proved very costly and troublesome to operate. France shut down its 1200-MWe demonstration breeder reactor in 1998 after it had operated at an average six-percent capacity over 13 years. Japan's 280-MWe demonstration breeder reactor first went critical in 1994 but was shut down by a sodium fire in 1995 and has not yet been brought back into operation. Russia's 600 MWe BN-600 demonstration reactor, in contrast, has been kept on line with an average capacity factor of about 74 percent since 1980 despite 15 sodium fires in 23 years. A follow-on demonstration reactor, the BN-800, has been intermittently under construction since 1986 and is currently again a high-priority project for Russia's nuclear establishment. India has begun to build a demonstration breeder reactor. But, despite a worldwide expenditure of perhaps $100 billion in current dollars thus far on developing and demonstrating breeder reactors with a total thermal capacity of about 9 GWt, no country has yet succeeded in commercializing them.

Commercial reprocessing has continued, however. Today civilian reprocessing on a large scale is underway in Britain, France, India and Russia and, in 2006, a large new reprocessing plant began operating in Japan.

Reprocessing has continued primarily because of a combination of local political pressures to do something about the problem of spent fuel accumulating at power-reactor sites and not-in-my-backyard political opposition elsewhere to geological repositories and central interim storage facilities for spent fuel. Indeed, Germany and Japan largely financed the French and U.K. multi-billion-dollar commercial reprocessing facilities as a way to export their spent-fuel storage problems, at least temporarily.

The respite was only temporary, however. After their reprocessing plants went into operation, Britain and France began to ship the solidified reprocessing waste back to the countries of origin – reopening the issue of where to store it. Germany's utilities finally decided to stop reprocessing, store newly generated spent fuel on site, and phase out nuclear power. Japan's nuclear utilities went a different route. They persuaded the rural prefecture of Aomori to store for 50 years the radioactive waste being returned from Europe in exchange for a large reprocessing plant and large tax payments to the local government.

Some countries, notably France and Germany, are recycling their separated plutonium in the form of mixed-oxide fuel (MOX) back into the reactors from whose spent fuel it was extracted. Japan plans to do the same but local government opposition has delayed this program for about a decade. The United Kingdom has been simply stockpiling its own separated civilian plutonium. Russia has been stockpiling the separated plutonium that it has recovered from the spent fuel of its own first-generation power reactors and those of Eastern Europe. As a result of these growing national stockpiles, the total global stock of separated civilian plutonium is about 250 tons.

The Bush Administration's Global Nuclear Energy Partnership

U.S. nuclear utilities too have been unable to ship their accumulating spent fuel off their reactor sites. The reason is delays in the licensing of the Department of Energy's (DOE) proposed Yucca Mountain, Nevada, geological repository that was supposed to have begun operations in 1998. The utilities have therefore been suing the DOE for the costs of building additional on-site dry-cask storage.

If the Yucca Mountain repository is licensed, the U.S. Department of Energy estimates its physical capacity as 105,000-200,000 tons. A recent study by the Electric Power Research Institute (EPRI) concludes that the capacity could be still higher -- from 260,000-570,000 tons. Current law, however, limits the quantity of spent fuel that can be stored there to 63,000 tons "until such a time as a second repository is in operation." U.S. reactors will have discharged this amount of spent fuel by 2008. In the spring of 2006, the DOE submitted legislation to Congress that would lift the legislated limit on the capacity of Yucca Mountain.

As an alternative option, in 2005, the U.S. Congress asked the Department of Energy to develop a plan for centralized interim storage and reprocessing of U.S. spent fuel. In May 2006, the DOE responded with a plan for a "Global Nuclear Energy Partnership" (GNEP). As illustrated below, it envisions building reprocessing plants that would separate spent light-water-reactor fuel into three streams: uranium, the 30-year-half-life fission products, strontium-90 and cesium-137; and plutonium mixed with the other transuranic elements neptunium, americium and curium.

FIGURE 5. Diagram showing Department of Energy's proposal for reprocessing U.S. spent fuel and fissioning the transuranics with fast-neutron reactors, shown here as Advanced Burner Reactors. After reprocessing, the 30-year half-life isotopes, cesium-137 and strontium-90, which dominate the radiological hazard until they decay away, would be placed in interim surface storage for some hundreds of years. This raises the question as to why the spent fuel should not be placed in such interim storage until the long-term future of nuclear power is clarified instead of rushing into a reprocessing and transmutation program that would ultimately cost more than $100 billion dollars to process the spent fuel from just the already existing U.S. power reactors.

The strontium-90 and cesium-137 would be resolidified and placed into interim surface storage for some hundreds of years. The transuranic elements would be recycled in fast-neutron reactors until they were fissioned. These are the same sodium-cooled reactors that previously were to be commercialized as plutonium breeder reactors. With the removal of the plutonium breeding uranium blankets around their cores, they now would be transuranic burner reactors. It is proposed that demonstration reprocessing and fast-neutron reactor plants be built and put into operation by 2020.

The purpose of this effort would be to drastically reduce the fraction of the long-lived radionuclides in the spent fuel going into the Yucca Mountain repository. This could decrease the long-term temperature increase of the mountain per ton of spent fuel and increase by up to one-hundredfold the amount of spent fuel that could be discharged from U.S. nuclear reactors before a new repository would have to be sited.

GNEP is controversial for two reasons: its cost and its impact on U.S. nonproliferation policy.

Cost. A 1996 U.S. National Academy of Sciences study estimated the extra cost of a separations and transmutation program for the first 62,000 tons of U.S. spent fuel, relative to the cost of simply storing the spent fuel in a repository, as "likely to be no less than $50 billion and easily could be over $100 billion." U.S. utilities, which have been paying the U.S. Government 0.1 cent per kilowatt-hour of nuclear-generated electricity for spent-fuel disposal services, have made clear that they are skeptical about the Department of Energy's proposed schedule and that they will not pay for the extra cost of building fast-neutron reactors.

The great cost of the DOE's proposed program and the fact that it proposes to store the most dangerous isotopes in the spent fuel on the surface for hundreds of years may eventually increase the appeal of interim storage without reprocessing.

Impact on U.S. nonproliferation policy. Following India's 1974 nuclear explosion, which used civilian plutonium separated with U.S. assistance, the United States reversed its policy of encouraging reprocessing and plutonium recycle worldwide. The U.S. policy became, in effect, "We don't reprocess and you don't need to either." Partly as a result, since 1974, only two additional countries have begun to reprocess, North Korea and Pakistan, both for weapons purposes. During the same period, Argentina, Belgium, Brazil, Germany and Italy, shut down their pilot reprocessing plants and South Korea and Taiwan abandoned their laboratory-scale reprocessing efforts.

The Department of Energy has responded in two ways to concerns that a new U.S. reprocessing initiative would undermine nonproliferation efforts:

It is developing reprocessing technologies that would not separate out pure plutonium. The proliferation-resistance of these technologies has been challenged, however, and the Argonne National Laboratory, which provides the technical analysis for DOE policy in this area, has responded by proposing ever more complex versions of its proposed UREX+ fuel cycle.
By proposing that enrichment and reprocessing be confined to "countries that already have substantial, well-established fuel cycles."

Indeed, the DOE named its proposed reprocessing and reycle program the Global Nuclear Energy Partnership to convey the idea that the United States and other countries with large nuclear programs would provide reprocessing services to other countries.

France, the United Kingdom and Russia have been doing this already but France and the United Kingdom have recently lost all of their foreign customers. Russia has kept a few because, unlike France and the United Kingdom, it is willing to keep other countries' plutonium and radioactive waste. In effect, it is providing permanent storage for foreign spent fuel, although with the fuel separated into three components: uranium, plutonium and high-level waste. Its clients are therefore happy for Russia to take their spent fuel, whether it reprocesses it or not. Indeed, the spent fuel from some first-generation VVER-440 reactors in Eastern Europe, Russia and Ukraine is reprocessed at the Mayak combine in the Urals while the spent fuel from some East European, Russian and Ukrainian VVER-1000 reactors is stored in a storage pool associated with an uncompleted reprocessing plant near Krasnoyarsk. Russia Federal Atomic Energy Agency (Rosatom) has indicated an interest in reprocessing -- or storing -- spent fuel from other countries as well. Recently, the Bush Administration has indicated its support for such a initiative.

Limit the Proliferation of National Fuel-Cycle Plants

Proposals to limit the proliferation of enrichment and reprocessing plants have been made periodically since the beginning of the nuclear era. The 1946 Acheson-Lillienthal report urged that such "sensitive facilities" should be placed under international ownership. After India used civilian plutonium to make a nuclear explosive device in 1974, there was a second wave of interest in limiting national ownership of reprocessing facilities, with studies launched in 1975, 1977, 1978, 1980, and 1987.

During the Cold War, the combination of the advanced nuclear states refusing to export fuel cycle facilities and the ability of the United States and Soviet Union to press their allied states not to develop such capabilities on their own was relatively effective. With the end of the Cold War, however, it become more difficult for Washington and Moscow to enforce nuclear abstinence. Also, over the past three decades a black-market developed for centrifuge plant designs and components outside government control. Efforts are therefore being made to strengthen control over technology exports and renewed proposals are being made for at least multinational -- if not international control of fuel-cycle facilities.

Strengthened technology export controls

In his talk at the National Defense University on February 11, 2004, President Bush called upon the Nuclear Suppliers Group to deny enrichment and reprocessing technologies "to any state that does not already possess full-scale, functioning enrichment and reprocessing plants," and, in compensation, ensure that states that do not have enrichment plants have reliable access to civilian reactor fuel. No member of the Nuclear Suppliers Group has contracted to export either type of plant to a non-weapons state other than Japan since the 1970s. However, there has been resistance to the proposal within the G-8 group of countries, which has been willing to support a formal moratorium on exports only on a year-by-year basis.

To deal with the problem of illicit technology exports exemplified by the A.Q. Khan network, the Bush Administration launched the Proliferation Security Initiative under which many countries have agreed to cooperate to intercept illicit shipments of weapon-of-mass-destruction related technologies such as gas centrifuges. Indeed, the interception of centrifuge components being shipped to Libya by the A.Q. Khan network is often cited as a model for the type of operation envisioned in the Proliferation Security Initiative.

The U.N. Security Council also passed in April 2004, UNSC resolution 1540, which requires all U.N. members to set up legal and regulatory systems to assure that "all States shall take and enforce effective measures to establish domestic controls to prevent the proliferation of nuclear, chemical, or biological weapons and their means of delivery."

Multinational control of fuel-cycle facilities

In his November 2003 speech to the United Nations General Assembly, IAEA Director General El Baradei proposed that enrichment and reprocessing be restricted "exclusively to facilities under multinational control." The Nuclear Suppliers Group guidelines already state that suppliers should "encourage" recipients to "accept, as an alternative to national plants, supplier involvement and/or other appropriate multinational participation in resulting facilities."

A subsequent study done for the IAEA by an expert group assessed a number of international and multinational approaches including the IAEA operating as "administrator of a fuel bank; promoting voluntary conversion of existing [fuel cycle] facilities to multinational nuclear arrangements ... and creating ... regional multinational nuclear arrangements for new facilities ..."

The panel observed, however, that "there is a consistent opposition by many [non-nuclear weapons states] to accept additional restrictions on their development of peaceful nuclear technology without equivalent progress on disarmament." Japanese, U.S. and enrichment-industry officials also expressed skepticism.

In January 2006, however, Russian President Putin suggested that Russia would be willing "to offer nuclear fuel cycle services, including enrichment under the control of the IAEA." The specifics of the proposal remain to be worked out. Russia has also offered to let Iran invest in a Russian enrichment facility as an alternative to building its own.

Two other ideas might be worth considering: the establishment of objective criteria for the ownership of national fuel-cycle facilities and a "black-box" approach to enrichment technology transfer:

Criteria for national ownership of fuel-cycle facilities

A criteria-based approach to national ownership of fuel-cycle facilities is apparently of interest to most of the members of the Nuclear Suppliers Group but is opposed by the United States. A Princeton University graduate workshop has proposed that the IAEA convene a conference to establish international agreement on objective criteria that would be have to be met before a country could qualify for hosting an enrichment plant. As a possible standard, it has suggested that a country have at least ten Gigawatts (GWe) of light-water reactor generating capacity, the equivalent of about ten full-sized power reactors. Supplying his much capacity with LEU would require about one million SWUs of enrichment work per year, enough to provide a domestic market for a Urenco-type enrichment plant large enough to be economically competitive with foreign enrichment services.

Such a criterion would disqualify all but four (Germany, South Korea, Japan, and Ukraine) of the 25 non-nuclear-weapon states with nuclear power reactors as well as several nuclear-weapon states. Two of these countries (Germany and Japan) already have enrichment plants, as do two of the countries below the threshold (Brazil and the Netherlands). An argument might be made, however, to exempt the Netherlands because it is part of the EU and its enrichment plant is part of an EU multinational company.

Black box enrichment plants

Accepting the current technological dominance of two centrifuge-enrichment suppliers, Urenco and Russia, would be another way to limit the proliferation of centrifuge technology. The danger of centrifuge technology proliferation would be reduced to the extent that other countries chose to import centrifuges rather than develop their own.

As already noted, this is happening already. France and the United States are building plants using imported Urenco centrifuges; and China uses Russian centrifuges. The Urenco contracts involve the export of its centrifuge technology only on a "black box" basis. The centrifuges are to be manufactured in the Netherlands and assembled by Urenco technicians in the recipient countries. Since the centrifuges are expected to operate for perhaps 20 years without maintenance, there is no need for the personnel of the host country to examine their interiors. Russia has made a similar black-box arrangement for the centrifuge plants that it has supplied to China.

Countries can, of course, choose their own technology, even if it is not economic. Japan, Brazil and Iran have opted for their own, less economical technologies or smaller plants with captive domestic markets. But each of these exceptions is special. Japan has a huge nuclear sector with 44 GWe of capacity; Brazil, launched its enrichment program when it had nuclear-weapons ambitions; and Iran is widely suspected of being interested in acquiring at least a nuclear-weapons option. And the fact that three weapon states are willing to acquire enrichment technology on a black-box basis should make such an approach appear less discriminatory to non-weapon states.