IPFM International Panel on Fissile Materials

[This entry is drawn from Chapter Three of the 2006 Global Fissile Material Report: "Production and Disposition of Fissile Materials." The printed version includes endnotes and, in some cases, additional figures. Entries are updated to reflect current data.]

Production and Disposition of Fissile Materials

Although the first uranium-enrichment plants were built to produce HEU and the first reactors were built to produce plutonium (both for weapons) globally, the civilian nuclear sector today vastly exceeds the nuclear-weapon sector in terms of the numbers of fuel cycle facilities and fissile-material production capabilities. Except in Russia, the dedicated production facilities built by the NPT nuclear weapon states to obtain both weapon-grade uranium and plutonium have ceased production and Russia’s facilities no longer produce fissile materials for weapons. India, Israel, North Korea and Pakistan are believed to be continuing to produce plutonium, and Pakistan, HEU for weapons.

In 1946, Robert Oppenheimer observed that, if there were a convention banning nuclear weapons:

"We know very well what we would do if we signed such a convention: We would not make atomic weapons, at least not to start with, but we would build enormous plants, and we would design these plants in such a way that they could be converted with the maximum ease and the minimum time delay to the production of atomic weapons saying, this is just in case somebody two-times us; we would stockpile uranium; we would keep as many of our developments secret as possible; we would locate our plants, not where they would do the most good for the production of power, but where they would do the most good for protection against enemy attack."

The Nuclear Fuel Cycle

All nuclear fuel cycles start today with uranium. Uranium is mined and the ore is milled to extract the uranium.

Natural uranium, which only contains 0.7 percent of U-235, is used directly as a fuel in a small fraction of the world’s power reactors. These are the heavy water reactors (HWRs or CANDU), developed by Canada but used today also in China, South Korea, India, Pakistan and Romania. The heavy water slows or 'moderates' the neutrons without absorbing them. Slow neutrons are preferentially absorbed on U-235 (250 times relative to U-238). As a result, it is possible to sustain a slow-neutron chain reaction in natural uranium despite the fact that only one atom in 140 is U-235. Very pure graphite was used to slow neutrons in the first plutonium-production reactors. The U.K.’s Magnox and AGR reactors, which use graphite as a moderator, are descended from its plutonium-production reactors.

Uranium can also be turned into uranium hexafluoride (UF6) to be enriched in the fraction of the chain-reacting isotope U-235. Most nuclear power reactors today are light water reactors (LWRs) that use ordinary water as moderator and coolant. Because ordinary water absorbs more neutrons than heavy water, they are designed to have less neutron moderation and therefore require fuel enriched to 3-5% U-235. The potential dual use of enrichment facilities manifests itself in the fact that they can be adapted to produce HEU for nuclear weapons.

In a reactor, fission produces heat that can be used to generate electricity while neutrons captured on U-238 in the fuel produce plutonium. About one percent of the U-238 in spent LWR fuel has been converted into plutonium.

After the fuel is discharged from a reactor, it is cooled in on-site pools for up to several years. The spent fuel can then either continue to be stored on site or elsewhere, or be reprocessed to recover the plutonium and uranium, with the fission products and other materials stored in tanks and then solidified as high-level waste. Ultimately, the spent fuel or high-level waste is to be stored in geological repositories. No repository is yet licensed or in operation but candidate sites are under development in the United States, Finland, and Sweden.

The separation of plutonium for civilian use was originally seen as a way to increase the energy that could be recovered from natural uranium, specifically from the U-238 isotope that makes up 99.3 percent of natural uranium. Conventional reactors are efficient only in fissioning the U-235 which makes up most of the remaining 0.7 percent.

The plan in most industrialized countries in the 1970s was that plutonium recovered from their spent LWR fuel would be used to provide the initial fuel for breeder reactors, that would then produce more plutonium from U-238 than they consumed, thus in effect, turning U-238 into their fuel. Breeder reactors have not matured as a safe and economic technology, however. As a result, some countries that reprocess spent fuel are storing their separated civilian plutonium, while others recycle it as fuel for LWRs. As noted earlier, almost any mixture of plutonium isotopes can be used in a nuclear weapon. Reprocessing therefore also is potentially a dual-use technology.

Uranium Isotope Separation

Since natural uranium contains only 0.7 percent of the chain-reacting U-235 and 99.3 percent non-chain-reacting U-238, it has to be enriched in the U-235 isotope to be usable in nuclear weapons. Natural uranium also has to be enriched in U-235 to be used in fuel for LWRs. However, as noted, for LWR reactor fuel, the U-235 needs to be enriched to only 3-5 percent, which is not weapon-usable.

The isotopes U-235 and U-238 are chemically virtually identical, differing in weight by only one percent. They are therefore very difficult to separate either chemically or physically. The ability to do so on a scale sufficient to make nuclear weapons or LWR fuel is possessed by only a relatively small number of nations.

In any enrichment facility, the process splits the feed (say natural uranium) into two streams: a product stream enriched in U-235, and a waste (or "tails") stream depleted in U-235. The work of isotope separation is measured in "separative work units" (SWUs). Likewise, the capacity of enrichment facilities is commonly described in SWU/yr.

To produce one kilogram of low-enriched uranium, with 4% U-235 for a LWR fuel takes about 7.5 kilograms of natural uranium feed and 6.5 SWU, if 0.2% U-235 is left in the depleted tails. To produce one kilogram of weapon-grade uranium (93% U-235) would take about 230 kilograms of natural uranium feed and 200 SWU, at a tails assay of 0.3%. Therefore, producing a kilogram of weapon-grade uranium requires about thirty times as much enrichment work as is required to produce a kilogram of LWR fuel. However, it takes about 20,000 kg a year of the low enriched uranium to fuel a typical 1000-Megawatt power reactor, as compared to the 25 kg of weapon-grade uranium to produce a nuclear weapon.

Therefore, even a small enrichment plant, such as the one that Iran proposes to build at Natanz, which is sized to fuel only a single power reactor, could make enough HEU for tens of bombs a year -- or if 20 tons of 4% LEU were fed into it, could produce enough weapon-grade uranium for four bombs in a little more than a week (see Table).

Today, two enrichment technologies are used on a commercial scale: gaseous diffusion and centrifuges. Gaseous diffusion plants remain operational in the United States and France, but both countries plan to switch to more economical gas centrifuge enrichment technology. For the same reason, all countries which have built new enrichment plants during the past three decades have chosen centrifuge technology.


Material and separative work required to fuel a 1000 MWe light-water reactor
Feed		Product		Separative Work	Time
150,000 kg	U(nat) at 0.71%	20,000 kg	LEU at 4% (Tails at 0.20%)	129,800 SWU	1 year
Material and separative work required to produce enough HEU for four bombs per year
Feed		Product		Separative Work	Time
150,000 kg	U(nat) at 0.71%	820 kg	HEU at 93% (Tails at 0.20%)	192,300 SWU
150,000 kg	U(nat) at 0.71%	100 kg	HEU at 93% (Tails at 0.65%)	14,200 SWU	40 days
20,000 kg	LEU at 4%	100 kg	HEU at 93% (Tails at 3.55%)	2,800 SWU	8 days

A 130.000 SWU/year enrichment plant could either supply a single 1000-MWe reactor or make weapon-grade uranium sufficient for many bombs. About 130,000 SWU are needed to produce the annual reloading of LEU fuel for a 1,000 MWe reactor. The same enrichment capacity could produce enough weapon-grade uranium for 26 nuclear weapons per year (assuming 25 kg of 93%-enriched uranium per weapon) or four weapons in 40 days. If the 20,000 kg of 4 percent enriched LEU produced for an annual reactor reload were instead recycled through the enrichment plant, it could be turned into enough HEU for 4 weapons in 8 days.


Country	Name/Location	Type	Status	Process	Capacity 1000s of SWUs/year
Brazil	Resende Enrichment	Civilian	Under Construction	GC	120
China	Lanzhou 2	Civilian	Under construction	GC	500
	Shaanxi Enrichment Plant	Civilian	In operation	GC	500
France	Eurodif (Georges Besse)	Civilian	In operation	GD	10800
	Georges Besse II	Civilian	Planned	GC	7500
Germany	Urenco Deutschland	Civilian	In operation	GC	1800 (4500)
India	Rattehalli	Military	In operation	GC	4-10
Iran	Natanz	Civilian	Under construction	GC	100-250
Japan	Rokkasho Enrichment Plant	Civilian	In operation	GC	1050
Netherlands	Urenco Nederland	Civilian	In operation	GC	2500 (3500)
Pakistan	Kahuta	Military	In operation	GC	15-20
Russia	Angarsk	Civilian	In operation	GC	1600
	Novouralsk (Sverdlovsk-44)	Civilian	In operation	GC	9800
	Zelenogorsk (Krasnoyarsk-45)	Civilian	In operation	GC	5800
	Seversk (Tomsk-7)	Civilian	In operation	GC	2800
U.K.	Capenhurst	Civilian	In operation	GC	4000
U.S.	Paducah Gaseous Diffusion	Civilian	In operation	GD	11000
	Portsmouth	Civilian	Standby	GD	7400
	Piketon, Ohio (USEC/DOE)	Civilian	Planned	GC	3500
	Eunice, NM (LES/Urenco)	Civilian	Planned	GC	3000

Large enrichment facilities, operational, under construction, and planned. Apart from some laboratory-scale facilities, all enrichment facilities today use either the gaseous diffusion (GD) or the gas centrifuge (GC) process. Since the large U.S. gaseous diffusion facility in Portsmouth, Ohio was shutdown in 2001, centrifuge facilities have accounted for more than half of global SWU production. Global enrichment demand is currently suppressed because a significant fraction of the LEU supply is provided by down-blended Russian HEU.

Uranium Isotope Separation: The Gas Centrifuge

Modern gas centrifuges spin uranium hexafluoride (UF6) gas at enormous speeds so that the uranium is pressed against the wall with more than 100,000 times the force of gravity. The molecules containing the heavier U-238 atoms concentrate slightly more toward the wall relative to the molecules containing the lighter U-235. Combined with an axial countercurrent circulation of the UF6 in the machine, this effect can be exploited to separate the two isotopes (see Figure below for an illustration).

Both throughput and enrichment achieved with a single machine are very small. The process is therefore repeated tens of times in a "cascade" of hundreds or thousands of centrifuges to produce uranium enriched to the 3-5 percent level used in most nuclear-power reactors. If the cascade is extended to three times as many stages or the uranium is recycled through the cascade three or four times, weapon-grade uranium can be produced.

The gas centrifuge for uranium enrichment and its large-scale use in an enrichment facility. The possibility of using centrifuges to separate isotopes was raised shortly after isotopes were discovered in 1919. The first experiments using centrifuges to separate isotopes of uranium (and other elements) were successfully carried out on a small scale prior to and during World War II, but the technology only became economically competitive in the 1970s. Today, centrifuges are the most economic enrichment technology, but also the most proliferation-prone.

Typical cascades for LEU and HEU production. The LEU cascade (top) produces uranium enriched to 4%, U-235 in 10 stages (seven processing enriched uranium and three depleted uranium); the HEU cascade (bottom) produces uranium enriched above 90% in 32 stages. Tails are 0.3% in both cases and the enrichment factor per stage is 1.3. The number of machines in both cascades is identical, but the HEU cascade produces much less product.

From a nonproliferation perspective, centrifuge technology has two major disadvantages relative to gas diffusion technology. First, the number of stages is much smaller (ten in the example in the figure above versus a thousand) and so the uranium moves through the cascade very quickly. The inventory held up in a typical cascade is more than a thousand metric tons in a gaseous diffusion plant as compared to a few kilograms in a centrifuge plant. This means that it could take only days to flush the uranium out of a centrifuge cascade and re-configure it for HEU production. This makes possible a "breakout" scenario, where peaceful technology is quickly converted to weapons use.

Second, clandestine centrifuge facilities are virtually impossible to detect with remote- sensing techniques. A centrifuge plant with a capacity to make HEU sufficient for a bomb or two per year could be small and indistinguishable from many other industrial buildings (see Figure below). Due to its low power consumption, there are no unusual thermal signatures as compared to other types of factories with comparable floor areas. Leakage of UF6 to the atmosphere from centrifuge facilities is also minimal because the gas in the pipes is below atmospheric pressure. Air therefore leaks into the centrifuges rather than the UF6 leaking out.

Centrifuge-containing buildings of “another country” (Pakistan’s Kahuta complex). These buildings were identified in a briefing on “Iran's Nuclear Fuel Cycle Facilities: A Pattern of Peaceful Intent?” distributed by the U.S. Government, September 2005. No distinguishing characteristics identify the highlighted centrifuge halls relative to the other buildings on the site. Satellite imagery from Google Earth; the rectangles are redrawn for clarity.

Plutonium Production

The weapon that destroyed Nagasaki contained six kilograms of plutonium. Plutonium does not occur naturally. It is produced in nuclear reactors when a U-238 nucleus absorbs a neutron creating short-lived U-239, which subsequently decays to neptunium and, ultimately, to plutonium (see Figure).

Making plutonium in a nuclear reactor. A neutron released by the fissioning of a chain-reacting U-235 nucleus is absorbed by the nucleus of a U-238 atom. The resulting U-239 nucleus decays with a half-life of 24 minutes into neptunium, which in turn decays into Pu-239.

Almost all reactors dedicated to the production of plutonium for weapons have been fueled with natural uranium. To avoid the buildup of unwanted heavier plutonium-isotopes (Pu-240, Pu-241, etc.) only about one seventh of the 0.7 percent U-235 in the fuel is fissioned. In such reactors, about 0.9 grams of plutonium are produced per gram of U-235 fissioned or, equivalently, per thermal megawatt day. For example, India's CIRUS research reactor, which has a thermal power of 40 megawatts, would, at a 70% capacity factor discharge annually about 10.2 tons of spent fuel containing about 9.2 kg of weapons grade plutonium.

Plutonium is also produced in civilian power reactors. In LWRs, the net plutonium production is only 0.2-0.3 grams of plutonium per thermal megawatt-day because about two thirds of the plutonium is fissioned in place during the long residency of the fuel in the reactor core. A 1000 MWe (3000 megawatt-thermal) LWR, operating at a 90-percent capacity factor produces about 250 kilograms of plutonium per year. Because the burn-up of the fuel is much higher than in production reactors, the fraction of heavier plutonium isotopes is more than 40 percent.

In the heavy-water-moderated CANDU power reactor, plutonium production is about twice as high as in LWRs and the fraction in the heavier plutonium isotopes is over 25 percent. CANDU reactors are continuously refueled instead of once every one or two years for LWRs, thus making safeguarding them more costly.

Several countries have pursued the development of fast-neutron or "plutonium-breeder" reactors. In breeder reactors, the reactor core is surrounded by a "blanket" of natural or depleted uranium that captures the neutrons escaping the core to make more plutonium. The plutonium that builds up in the blanket is weapon-grade.

Thus, uranium-based spent fuel from all types of reactors will contain substantial amounts of plutonium. As long as the plutonium remains embedded in the spent fuel along with the highly radioactive fission products, however, it is relatively inaccessible. Spent fuel can only be handled remotely due to the very intense radiation field, which makes its diversion or theft a rather unrealistic scenario. Separating the plutonium from the fission products and uranium therefore makes diversion or theft a much greater concern. Separated plutonium can be handled without radiation shielding. It is dangerous only when inhaled or ingested.

Separation of the plutonium has to be done in a "reprocessing" operation. With the current PUREX technology, the spent fuel is chopped into small pieces, and dissolved in hot nitric acid. The plutonium is extracted in an organic solvent which is mixed with the nitric acid using blenders and pulse columns, and then separated with centrifuge extractors. Because all of this has to be done behind heavy shielding and with remote handling, reprocessing requires both resources and technical experience. However, detailed descriptions of the process have been available in technical literature since the 1950s.

Military reprocessing. All of the nuclear-weapon states have produced plutonium through reprocessing. As indicated in the previous chapter, the United States, United Kingdom, France, and China have stopped producing plutonium for weapons. Russia continues to produce about 1.2 tons of separated plutonium a year as an unwanted byproduct of the continued operation of three of its plutonium-production reactors. Israel, India, Pakistan, and North Korea have not indicated that they have stopped plutonium production for weapons.

Civil plutonium produced in commercial reactors. The figure below shows the growth of nuclear power worldwide. At present, the world nuclear capacity stands at about 370 gigawatts-electric (GWe), approximately 87 percent of which is in LWRs. The total spent fuel generated annually is approximately 7000 metric tons, containing about 70 metric tons of plutonium. The cumulative total of plutonium in spent fuel worldwide at the end of 2005 was approximately 1450 metric tons. Roughly one-third of the spent fuel generated each year is reprocessed; the remainder is being stored at reactor sites.

Nuclear power capacity, historically. Global nuclear capacity grew rapidly during the 1970s and 1980s. Public opposition, high costs, unresolved waste issues, and in particular also the accidents of Three-Mile-Island and Chernobyl in 1979 and 1986, led to a sharp decline of new orders of nuclear power plants worldwide. The installed capacity has reached 370 GW(e) in 2006, equivalent to about 370 power reactors. Due to the shutdown of aging reactors, this capacity will decline during the next few decades unless the rate of ordering new nuclear power plants rises above an average of about ten per year.

Civilian reprocessing. Reprocessing of civilian spent fuel is being done at present in the United Kingdom, France, Russia, India, and Japan (see Table). This civilian separation of plutonium stemmed originally from the historical interest of some industrialized countries in commercializing plutonium-breeder reactors. This interest, which peaked in the 1970s, was driven by an expectation that the world's nuclear generating capacity would grow to thousands of gigawatts by the year 2000 and approach 10,000 GWe in 2020. Such a huge capacity could not have been supported by known reserves of high-grade uranium ore.


COUNTRY	NAME/LOCATION	TYPE	STATUS	DESIGN CAPACITY [tHM/yr]
France	La Hague - UP2	Civilian	In operation	1000
	La Hague - UP3	Civilian	In operation	1000
India	Trombay	Military	In operation	50
	Tarapur	(unclear)	In operation	100
	Kalpakkam	(unclear)	In operation	100
Israel	Dimona	Military	In operation	40-100
Japan	JNC Tokai Reprocessing Plant	Civilian	In operation	210
	Rokkasho Reprocessing Plant	Civilian	Under construction	800
Pakistan	Nilore	Military	In operation	10-20
Russia	RT-1 Ozersk (Mayak)	Civilian	In operation	400
	RT-2, Krasnoyarsk, 1st Line	Civilian	Deferred	800
U.K.	BNFL B205	Civilian	In operation	1500
	BNFL Thorp	Civilian	Operation currently suspended	900

Reprocessing facilities worldwide, operational and under construction. As listed by the IAEA's Nuclear Fuel Cycle Information System (NFCIS, data retrieved in February 2006), except where indicated. Actual throughput in reprocessing plants is often a small fraction of the design capacity.

Efforts to commercialize plutonium breeder reactors have largely failed because of their poor economics and technical difficulties. A few countries have begun using their separated plutonium to make mixed-oxide (MOX, uranium-plutonium) fuel for conventional light-water reactors as a substitute for standard LEU fuel. MOX use has been limited so far to only a few countries in Western Europe, however. The United Kingdom and Russia are simply storing their separated plutonium and Japan has not yet overcome local opposition to MOX fuel. As a result, the global stockpile of separated civilian plutonium has been growing steadily for decades. The figure below illustrates this trend, going back to 1996, when all countries with stocks of civilian separated plutonium except India, started to officially declare their civilian plutonium holdings to the IAEA. With Japan's new reprocessing plant going into operation in 2006, the growth of the global stockpile of separated civilian plutonium will continue for some time, even if the United Kingdom ends its reprocessing operations by 2012, as currently planned.

The United States abandoned reprocessing in the late 1970s for nonproliferation and economic reasons. Recently, however, the Bush Administration embraced reprocessing as part of its proposed Global Nuclear Energy Partnership (GNEP), the idea of recycling plutonium and other transuranic isotopes in fast-neutron reactors.

Stockpiles of civilian separated plutonium are growing. Since 1996, the civilian plutonium stockpile has increased by more than 80 metric tons and exceeded 240 metric tons at the end of 2004. This total does not include the nearly 100 tons of weapon-grade plutonium declared excess by the United States and Russia. Japan, Germany and some other smaller West European countries store their plutonium at the French and U.K. reprocessing plants until it can be used. This practice increases the inventories attributed here to France and the U.K. The civilian stockpile of separated plutonium is likely to continue to grow rapidly because of Japan’s large new reprocessing plant at Rokkasho, which became operational in 2006, although Japan has not yet been able to recycle MOX in any of the 40 tons of separated plutonium that it has already accumulated. Data from IAEA INFCIRC/549 declarations.

Disposition of Fissile Materials

Highly enriched uranium. From a technical perspective, the disposition of HEU is simple and straightforward. It can be down-blended to lower enrichment by mixing with depleted, natural, or slightly-enriched uranium. This process cannot be reversed without re-enrichment. It is also economically attractive since the LEU product can be sold for use as commercial reactor fuel at a price several times higher than the cost of the blend down process.

Russia agreed to sell 500 tons of its excess weapon-grade HEU, after down-blending to LEU, to the United States in a groundbreaking 1993 bilateral agreement. The rate of blend down is limited to 30 tons per year, however, so as not to disrupt the uranium and enrichment markets. The United States is similarly down-blending most of the 198 tons of HEU that it has declared excess for military purposes.

Plutonium. The debate on the management of separated plutonium inventories has been primarily focused on the weapons plutonium declared excess by the United States and Russia. Most of the considerations are equally applicable, however, to the disposition of civilian stocks of separated plutonium that are accumulating in Europe, Russia and soon Japan.

Two approaches are currently being pursued:

Consolidating and storing excess inventories indefinitely in high-security facilities such as that built at Mayak for excess Russian weapons plutonium, with U.S. funds. This approach is only as effective as the institution responsible for security.
Mixing the plutonium with fission products – either through irradiation or directly – so as to recreate the radiation barrier that was eliminated when the plutonium was separated. This concept is sometimes measured by the "spent fuel standard," which was defined in the National Academy studies as the objective of making excess plutonium "roughly as inaccessible for weapons use as the much larger and growing stock of plutonium in spent fuel." One way to do this is by mixing the plutonium with uranium to make mixed oxide fuel and then irradiating the fuel in power reactors. MOX fuel containing about four percent weapon-grade plutonium mixed with depleted uranium can be used as an alternative to LEU fuel in LWR. In a second approach, the plutonium would be mixed with already existing fission products in highly radioactive reprocessing waste – or with spent fuel, to create a radiological barrier.

In the long term (after a century or so of cooling), the gamma-radiation field around spent fuel will die down to levels that are no longer considered adequate for self protection and additional barriers such as deep safeguarded underground storage would be required.

Russia and the United States agreed in 2000 to eliminate 34 tons of weapons plutonium each. Russia agreed, however, only on the conditions that its plutonium and most of the U.S. plutonium be disposed of in MOX and that other governments fund the building and operation of the necessary infrastructure in Russia. Progress has been stalled for years, however, by disagreements between the United States and Russia with regard to immunity from liability of U.S. contractors in Russia. The G-7 governments have committed $800 million, but that is nowhere near enough to cover both construction and operation of a MOX-fuel fabrication plant. The estimated cost of constructing the U.S. MOX facility increased from less than $1 billion to $3.5 billion between 2002 and 2005. In any case, Russia would prefer to use the assistance to help it build a plutonium breeder reactor and use it to irradiate the plutonium. In 2006, the U.S. Congress began to reassess this program, including considering decoupling the U.S. and Russian plutonium disposition programs and shifting the focus of the U.S. plutonium-disposition program to the less costly option of immobilizing the plutonium with fission products.

Currently, neither Russia nor the United Kingdom has definite plans for how to dispose their excess stocks of civilian plutonium. Japan plans to dispose of its stock via recycle in MOX in light water reactors but has not yet begun because of public opposition.