Frank von Hippel

The possibility of a plutonium-fueled nuclear reactor that could produce more fuel than it consumed (a "breeder reactor") was first raised during World War II in the United States by scientists in the atomic bomb program. In the following two decades, the Soviet Union, the United Kingdom, France, Germany, Japan and India followed the United States in establishing national plutonium breeder reactor programs, while Belgium, Italy and the Netherlands joined the French and German programs as partners. In all of these programs, the main driver was the hope of solving the long-term energy supply problem using the large scale deployment of nuclear energy for electric power. Plutonium-fueled breeder reactors appeared to offer a way to avoid a potential shortage of the low-cost uranium required to support such an ambitious vision using other kinds of reactors.

Uranium proved to be much more abundant than originally imagined and, after a fast start, nuclear power growth slowed dramatically in the late 1980s and global nuclear capacity is today about one-tenth the level that had been projected in the early 1970s. The urgency of deploying fast-neutron reactors for plutonium breeding therefore abated — at least in the western Organization for Economic Co-operation and Development (OECD) countries. In India and Russia, however, concerns about potential near-term uranium shortages persist, and new demonstration breeder reactors are being built. China, which currently is building up its nuclear capacity at an enormous rate, is considering the possibility of building two Russian-designed breeder reactors. Because of the high costs and reliability and safety issues that are detailed below, however, no commercial breeder reactors have been deployed.

Interest in fast-neutron reactors persists in the OECD countries for a new reason, political difficulties with storing or disposing of spent fuel. "Reprocessing" spent fuel does not eliminate the problem of siting a geological repository but a reprocessing plant does provide an interim destination that has proved a path forward with regard to the spent fuel problem in a number of nations.

Spent-fuel reprocessing was originally launched in countries that planned to deploy breeder reactors. They wanted separated plutonium for manufacturing startup fuel for their first breeder reactors. Standard light-water-reactor spent fuel contains about one percent plutonium. In the absence of breeder reactors, the separated plutonium has become a disposal problem and some countries have decided to recycle it into fuel for the same reactors that produced it. Slow-neutron reactors are relatively ineffective, however, in fissioning some of the plutonium isotopes, which therefore build up in recycled fuel.

Fast-neutron-reactor advocates argue that, if the plutonium and other long — lived transuranics in spent fuel could be fissioned almost entirely, the political problem of finding a geological disposal site for radioactive waste consisting of mostly shorter-lived fission products would become much easier. Fast neutron reactors would be more effective in fissioning all the transuranic isotopes. Fast — neutron breeder reactors could be converted into transuranic "burner" reactors by removing the plutonium-breeding uranium blankets around their cores and flattening the cores into more of a "pancake" shape so that more neutrons would leak out of them.

	MWe	MWt	Operation
France
Rapsodie		40	1967-83
Phenix	250		1973-2009
Superphenix	1240		1985-98
India
FBTR		40	1985-
PFBR	500		2010?
japan
Joyo		140	1977-
Monju	280		1994-95, 2010?
USSR/Russia
BR-5		5	1959-2004
BOR-60	12		1969-

	MWe	MWt	Operation
USSR/Russia (cont.)
BN-350 (Kazakhstan)	350		1972-99
BN-600	600		1980-
BN-800	800		2014?
United Kingdom
DFR	15		1959-77
PFR	250		1974-94
United States
EBR-I	0.2		1951-63
EBR-II	20		1963-94
Fermi 1	66		1963-72
SEFOR		20	1969-72
Fast Flux Test Facility		400	1980-93

Table 1.1 Major experimental, pilot and demonstration fast breeder reactors.1

This report looks at the experience and status of breeder reactor programs in France, India, Japan, the Soviet Union/Russia, the United Kingdom and the United States. The major breeder reactors built in these countries are listed in table 1.1. Germany also built two breeder reactors. All were sodium cooled.

The problems described in the country case studies in the following chapters make it hard to dispute Admiral Hyman Rickover’s summation in 1956, based on his experience with a sodium-cooled reactor developed to power an early U. S. nuclear submarine, that such reactors are "expensive to build, complex to operate, susceptible to prolonged shutdown as a result of even minor malfunctions, and difficult and time-consuming to repair."2

Fast-neutron breeder reactors

Fissile isotopes are the essential nuclear materials in both nuclear reactors and nuclear weapons. They undergo fission when they absorb neutrons and, on average, release more neutrons than they absorb. This makes a sustained chain — reaction possible in a "supercritical mass." This supercritical mass must contain a significant concentration of fissile isotopes and must be large enough so that only a small fraction of the neutrons escape without interacting.

The most important fissile materials are uranium-235 and plutonium-239. Uranium-235 is found in nature, constituting 0.7 percent of natural uranium. Plutonium-239 is created when uranium-238 (99.3 percent of natural uranium) absorbs a neutron (figure 1.1).

Figure 1.1 Plutonium breeding. A plutonium breeder reactor produces more plutonium than it consumes by using its extra fission neutrons to convert uranium-238 to uranium-239, which changes by radioactive decays involving electron and neutrino emission into neptunium-239 and then plutonium-239.

The vast majority of deployed power reactors around the world are fueled with low-enriched uranium and use a neutron "moderator" — in most cases, ordinary water, which also serves as the reactor coolant — that slows the neutrons and increases the likelihood that they will be captured by uranium-235 and cause it to fission. Such reactors are called "light-water reactors" and are fueled by uranium typically enriched to four-five percent in uranium-235. Light-water reactors are so named to distinguish them from the "heavy-water reactors" developed by Canada, which are fueled by natural uranium. In both types of reactors, the neutrons lose most of their energy in collisions with hydrogen, ordinary hydrogen in light — water reactors and heavy-hydrogen or deuterium in heavy-water reactors. In both types of reactors, some of the extra neutrons from uranium-235 fissions are also captured by uranium-238, converting it into chain-reacting plutonium-239 — but not enough to replace the fissioned uranium-235.

Virtually all breeder reactor programs have focused on reactors that do not use water as a coolant, so that the neutrons propagating the chain-reaction remain energetic (fast).

In order to be supercritical with fast neutrons, the "cores" of breeder reactors contain over 20 percent of fissile material — usually plutonium-239 — mixed with natural or "depleted uranium" (the residue after uranium-235 has been extracted from natural uranium by uranium-enrichment plants). Surrounding this core on all sides is a "blanket" — usually also consisting of natural or depleted uranium. Neutrons that leak out of the core are absorbed by the uranium-238 in the blanket and convert it into plutonium. Because such a reactor produces more plutonium than it consumes, its ultimate fuel is uranium-238, which is 140 times more abundant than uranium-235.

Plutonium breeder reactor programs have focused on fast-neutron reactors because, when a fast neutron fissions a plutonium-239 nucleus, more secondary neutrons are produced per fission than with any other combination of neutron speed and fissile isotope.3 Fast-neutron plutonium-fueled reactors can therefore breed extra fissile material more rapidly than any other reactor system. Despite the safety, cost and reliability issues of fast-neutron reactors, this fact determined their choice as the preferred technology at a time when the global population of nuclear power reactors was expected to double every decade indefinitely. The extra plutonium produced by fast-neutron reactors could be used to provide startup fuel for additional plutonium breeder reactors, allowing the number of breeder reactors to grow at a high rate.

In Russia, given the cost and safety problems associated with plutonium fuel thus far, demonstration fast-neutron reactors have been fueled with highly-enriched uranium, enriched to between 20 and 30 percent uranium-235.

Why commercialization of breeder reactors failed

The rationale for pursuing breeder reactors — sometimes explicit and sometimes implicit — was based on the following key assumptions:

1. Uranium is scarce and high-grade deposits would quickly become depleted if fission power were deployed on a large scale;

2. Breeder reactors would quickly become economically competitive with the light-water reactors that dominate nuclear power today;

3. Breeder reactors could be as safe and reliable as light-water reactors; and,

4. The proliferation risks posed by breeders and their "closed" fuel cycle, in which plutonium would be recycled, could be managed.

Each of these assumptions has proven to be wrong.

Uranium is cheap and abundant. Breeder reactors were seen as a solution for the uranium scarcity problem because, by converting uranium-238 into chain­reacting plutonium, they can potentially increase one-hundred-fold the amount of fission energy that can be extracted from a kilogram (kg) of uranium and make it economically feasible to mine much lower grades of uranium ore.4

In 2007, uranium requirements for the global fleet of nuclear power reactors were 67,000 metric tons — approximately 180 tons per gigawatt of generating capacity per year. The International Atomic Energy Agency (IAEA) projects that global nuclear capacity will increase and that uranium requirements will increase correspondingly to between 94,000 and 122,000 tons a year in 2030.5

In 2008, the biennial report put out by the OECD Nuclear Energy Agency, Uranium 2007: Resources, Production and Demand — also known as "the Red Book" — found that, despite inflation, global known conventional resources of uranium recoverable for less than $130/kg had increased from about 4.7 to about 5.5 million tons. The Red Book also reported estimates from 27 countries that, with further exploration, an additional 7.6 million tons of uranium would be discovered in the same cost range.6 At $130/kg, the cost of uranium would contribute 0.3 U. S. cents to the cost of a kilowatt-hour of nuclear electricity.

In the long run, worldwide, the amount of uranium recoverable at low cost is virtually certain to be far greater than the numbers reported in the Red Book. If plausible estimates of geological abundance are used, the amount of uranium still to be discovered at recovery costs up to $130/kg would be 50-126 million tons.7 This corresponds to 500 to 1000 times the projected demand in 2030.

It will be seen from figure 1.2 that the price of uranium on the spot market went significantly above $130/kg during the late 1970s and then again after 2005. Except for these two periods when there was disequilibrium between supply and demand, prices have been less than $50 per kg. The 1970s price peak was due to the expectation of an enormous expansion in nuclear power capacity. This expectation was not realized but large stockpiles of uranium were built up and then sold off during the subsequent decades resulting in the closure of many uranium mines. The sale by Russia to the U. S. of low-enriched uranium blended down from 500 tons of weapon-grade uranium from excess Cold War weapons at a rate sufficient to fuel half of the U. S. nuclear capacity extended the period of low demand for freshly mined uranium.8 The stockpiles of natural uranium have been largely used up, however, and the blend-down of the Russian weapon-grade uranium will be completed in 2013. The most recent uranium price peak therefore reflected, at least in part, the expectation, compounded by speculation, that there might be uranium shortages before uranium-mining capacity increases again to the level required to support growing demand.

In any case, unlike the situation with oil or gas-fueled power plants, the cost of uranium fuel can double without having a significant impact on the cost of nuclear power. As noted above, at $130/kg, the cost of uranium contributes only 0.3 cents to the cost of a kilowatt-hour (kWh), which is about 5 percent of the cost of electricity produced by a new light-water reactor.9

Breeder reactors are costly to build and operate. Governments of countries in the OECD have together reported that they have spent about $50 billion (2007$) on fission and breeder reactor research and development (figure 1.3). Of this total, the United States reported that it had spent $15 billion, Japan $12 billion, the United Kingdom $8 billion, Germany $6 billion and Italy $5 billion. France

Figure 1.2 History of the price of uranium since 1970.10

Figure 1.3 Total fission and breeder research, development and demonstration funding in the OECD countries that had substantial breeder programs (1974 to 2007);11 Belgium, France, Germany, Italy, Japan, Netherlands, the United Kingdom and the United States.

reported only $1 billion in expenditures but this was obviously an incomplete report, given that the total cost of the Superphenix demonstration project alone is estimated at FRF 65 billion (1998FRF) or $14 billion (2007$) (see chapter 2).

Russia and India, which are both outside the OECD, have spent large amounts on breeder research, development and demonstration. The Soviet Union and Russia alone have spent an estimated $12 billion (see chapter 5). Yet none of these efforts has produced a reactor that is anywhere near economically competitive with light-water reactors.

The individual country studies make clear that, without astronomically high uranium prices, breeder reactors are unlikely to be economically competitive with light-water reactors. For "demonstration" liquid-sodium-cooled reactors the capital costs per kilowatt (KW) generating capacity have typically been more than twice those of water-cooled reactors of comparable capacity. Since breeder reactors were never built in quantity, it could be expected that, in production, this cost ratio would decline. Few if any argue today, however, that the capital costs for breeder reactors could be less than 25 percent higher than for water-cooled reactors of similar generating capacities. This would be a capital cost difference on the order of $1000 per kilowatt of generating capacity. At a 10 percent capital charge and a 90 percent average capacity factor, this would translate to a cost difference of about 1.3 cents per kilowatt hour.

Detailed economic comparisons of light-water reactors and breeder reactors using different breeding ratios, fuel reprocessing and fabrication costs, and capital costs show that direct disposal of spent light-water-reactor fuel would be far less expensive than reprocessing and plutonium recycle in breeder reactors under a wide range of assumptions.12

Fast-neutron reactors have special safety problems. As already noted, fast-neutron reactors cannot use water as a coolant because collisions with the hydrogen nuclei in water quickly remove most of the kinetic energy from the neutrons. Also, in order to sustain a chain-reaction with fast neutrons, the fissile material in a reactor core must be more concentrated. As a result, fast-neutron — reactor cores are smaller than those of light-water reactors with the same power. This necessitates the use of a coolant that can efficiently carry away the heat. The coolant that has been used in all demonstration breeder reactors to date is a liquid metal that melts at relatively low temperatures — sodium.

Sodium has both safety advantages and disadvantages compared to water. Its primary safety advantage is that the reactor operates below the boiling point of liquid sodium (883 °С) and therefore at low pressure. By contrast, water-cooled reactors operate at high pressures — over 150 atmospheres for pressurized water reactors. Therefore, if there is a large break in a pipe of a water-cooled reactor, the water flashes into steam, leaving the reactor’s intensely hot fuel without coolant unless the core is flooded with emergency cooling water. In the case of a sodium — cooled reactor, however, unless the break is below the top of the core, the sodium will continue to cover the core and absorb heat.

Sodium’s major disadvantage is that it reacts violently with water and burns if exposed to air. The steam generators, in which molten-sodium and high-pressure water are separated by thin metal, have proved to be one of the most troublesome features of breeder reactors. Any leak results in a reaction that can rupture the tubes and lead to a major sodium-water fire.

As the country studies detail, a large fraction of the liquid-sodium-cooled reactors that have been built have been shut down for long periods by sodium fires. Russia’s BN-350 had a huge sodium fire. The follow-on BN-600 reactor was designed with its steam generators in separate bunkers to contain sodium-water fires and with an extra steam generator so a fire-damaged steam generator can be repaired while the reactor continues to operate using the extra steam generator. Between 1980 and 1997, the BN-600 had 27 sodium leaks, 14 of which resulted in sodium fires (see chapter 5).

Leaks from pipes into the air have also resulted in serious fires. In 1995, Japan’s prototype fast reactor, Monju, experienced a major sodium-air fire. Restart has been repeatedly delayed, and, as of the end of 2009, the reactor was still shut down. France’s Rapsodie, Phinix and Superphinix breeder reactors and the UK’s Dounreay Fast Reactor (DFR) and Prototype Fast Reactor (PFR) all suffered significant sodium leaks, some of which resulted in serious fires.

Sodium also creates radiation problems. When it absorbs a neutron, ordinary sodium-23 becomes sodium-24, a gamma-emitting isotope with a 15-hour half­life. The sodium that cools the core therefore becomes intensely radioactive. To ensure that a steam-generator fire does not disperse radioactive sodium, reactor designers have inserted an intermediate sodium loop. The heat generated from the reactor is transferred to non-radioactive sodium through a sodium-sodium heat exchanger. The non-radioactive sodium delivers the heat to the steam generators. The extra sodium loops and associated pumps contribute to the high capital costs of breeder reactors.

Finally, light-water-cooled reactors have the critical safety characteristic that, if the water moderator is lost, the chain-reaction stops. It is impossible to sustain a chain-reaction in 4 to 5 percent enriched uranium without slowing the neutrons so that they are captured preferentially by uranium-235. In the absence of the water, the fast neutrons will be absorbed mostly in uranium-238 and the chain — reaction ends.

By contrast, in a fast-neutron reactor, the concentration of plutonium is high enough that it can sustain a chain-reaction even in the event of a coolant loss. Indeed, except for special core configurations, the reactivity will increase if the coolant is lost.13 Furthermore, if the core heats up to the point of collapse, it can assume a more critical configuration and blow itself apart in a small nuclear explosion.14 Whether such an explosive core disassembly could release enough energy to rupture a reactor containment and cause a Chernobyl-scale release of radioactivity into the environment is a major concern and subject of debate. (See chapter 3 for a discussion of this debate in India.)

Sodium-cooled reactors have severe reliability problems. The reliability of light-water reactors has increased to the point where, on average, they operate at about 80 percent of their generating capacity. By contrast, a large fraction of sodium-cooled demonstration reactors have been shut down most of the time that they should have been generating electric power. A significant part of the problem has been the difficulty of maintaining and repairing the reactor hardware that is immersed in sodium. The requirement to keep air from coming into contact with sodium makes refueling and repairs inside the reactor vessel more complicated and lengthy than for water-cooled reactors. During repairs, the fuel has to be removed, the sodium drained and the entire system flushed carefully to remove residual sodium without causing an explosion. Such preparations can take months or years.

In contrast, when a water-cooled reactor is shut down, the top of the pressure vessel can be removed and the reactor cavity that holds the pressure vessel can be flooded with water to provide shielding against the radioactivity of the fuel and the irradiated steel. Repairs can take place guided by underwater periscopes and video cameras.

The history of the world’s only commercial-sized breeder reactor, France’s Superphenix, is dominated by lengthy shutdowns for repairs (see chapter 2). Superphenix went critical and was connected to the grid in January 1986 but was shut down more than half of the time until operations ceased in December 1996. Its lifetime capacity factor — the ratio of the number of kilowatt-hours that it generated to the number it could have generated had it operated continually at full capacity — was less than 7 percent. The histories of Japan’s Monju and the U. K.’s Dounreay and Prototype Fast Reactors and the U. S. Enrico Fermi 1 demonstration breeder reactor power plants were similarly characterized by prolonged shutdowns (see chapters 4, 6 and 7). Russia’s BN-600 has experienced a respectable capacity factor but only because of the willingness of its operators to continue to operate it despite multiple sodium fires.

The fast-neutron reactor fuel cycle provides easy access to plutonium for weapons. All reactors produce plutonium in their fuel but breeder reactors require plutonium recycle, the separation of plutonium from the ferociously radioactive fission products in the spent fuel. This makes the plutonium more accessible to would-be nuclear-weapon makers. Breeder reactors — and separation of plutonium from the spent fuel of ordinary reactors to provide startup fuel for breeder reactors — therefore create proliferation problems.

This fact became dramatically clear in 1974, when India used the first plutonium separated for its breeder reactor program to make a "peaceful nuclear explosion." Breeders themselves have also been used to produce plutonium for weapons. France used its Phenix breeder reactor to make weapon-grade plutonium in its blanket. India, by refusing to place its breeder reactors under international safeguards as part of the U. S.-India nuclear deal, has raised concerns that it might do the same.

India’s Prototype Fast Breeder Reactor (PFBR), expected to be completed in 2010, will have the capacity to make 90 kg of weapon-grade plutonium per year, if only the radial blanket is reprocessed separately and 140 kg per year if both radial and axial blankets are reprocessed.15 The Nagasaki bomb contained 6 kg of weapon-grade plutonium and modern weapons designs contain less. At 5 kg per warhead, the PFBR would produce enough weapon-grade plutonium for 20-30 nuclear weapons a year, a huge increase in production capacity in the context of the South Asian nuclear arms race.

The G. W. Bush Administration proposed to make reprocessing more "proliferation resistant" by leaving some of the other transuranic elements (neptunium, americium and curium) mixed with the plutonium.16 Even if all the transuranics

Figure 1.4 Dose rate of a 4.4 kg container of various mixtures of separated transuranics compared to spent fuel. Peak dose rates only approach the IAEA’s self-protection standard (100 rem/hour at one meter distance) if high-activity fission products are included. Cesium-137, which has a half-life of 30 years, dominates the radiation field from spent fuel after ten years. (Weapon-grade plutonium [WPu], reactor-grade plutonium [RPu], neptunium [Np], americium [Am], transuranic waste [TRU], cesium [Cs] and strontium [Sr]).17

were left mixed with the plutonium, however — a project that the U. S. Department of Energy abandoned when it learned that the technology was not in hand — the gamma radiation field surrounding the mix would still be less than one-hundredth the level the IAEA considers self-protecting against theft and thousands of times less than the radiation field surrounding plutonium when it is in spent fuel (figure 1.4).