DOI: 10.2968/066003007
Bulletin
of the
Atomic
Scientists
It’s time to give up on breeder
reactors
Since the dawn of the nuclear age, nuclear
energy advocates have dreamed of a
reactor that could produce more fuel than it
used. more than 60 years and $100 billion
later, that vision remains as far from reality
as ever.
By ThOmaS B. COChran, harOlD a. FeIveSOn, ZIa mIan, m. v. ramana,
myCle SChneIDer & Frank n. vOn hIppel
T
he possibility of a plutonium-fueled nuclear
reactor that could produce more plutonium than it consumed (dubbed a “breeder reactor”) was first raised during World War II by scientists in the U.S. atomic bomb
program. They were concerned that uranium 235, the rare chainreacting isotope that fuels today’s nuclear reactors, was insufficiently abundant on Earth to support a large-scale deployment of
nuclear power. Over the next 20 years, Britain, France, Germany,
India, Japan, and the Soviet Union followed the United States in establishing national plutonium breeder reactor programs. (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 by deploying large
numbers of nuclear power reactors.
In “Fast Breeder Reactor Programs: History and Status,” a new
report by the International Panel on Fissile Materials, experiences with fast breeder reactors in six countries are examined.1 These
studies make clear that the assumptions driving the pursuit of
breeder reactors for the past six decades have proven to be wrong.
False assumptions. The rationale for pursuing breeder reactors
was based on the following key assumptions (sometimes explicit,
sometimes implicit): (1) Uranium is scarce, and high-grade deposits would quickly become depleted if light water nuclear reactors,
which do not breed more fuel than they consume, were deployed on
a large scale; (2) breeder reactors would quickly become economically competitive with light water reactors (the dominant reactor
IT IS 6 MINUTES TO MIDNIGHT
www.thebulletin.org
®
Bulletin of the Atomic ScientiStS | WWW.theBulletin.oRG may /j u n e 2 0 1 0
50
design used 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.
These assumptions, however, all proved to be wrong.
Uranium is cheap and abundant. The Nuclear Energy Agency’s Uranium 2007: Resources, Production, and Demand report—
popularly known as the Red Book—found that known global conventional resources of uranium recoverable for less than $130 per
kilogram amounted to approximately 5.5 million tons and estimated
that, with further exploration, an additional 7.6 million tons would
be discovered in the same cost range. (Historically, the spot price of
uranium has always been less than $100 per kilogram, except for a
period in the late 1970s and around 2007.)
In 2007, uranium demand for the global fleet of nuclear power
reactors was 67,000 metric tons—approximately 180 tons per gigawatt of generating capacity per year. The International Atomic Energy Agency (IAEA) projects that demand for uranium could increase to between 94,000 and 122,000 tons per year by 2030 as part
of an assumed global increase in nuclear energy capacity.2 Thus, at
the highest rate of consumption projected for 2030, estimated global uranium resources would last for about 100 years. But even this
projection is most likely conservative.
Plausible estimates of geological abundance predict that the
amount of uranium recoverable at up to $130 per kilogram is virtually certain to be far greater than the numbers reported in the
Red Book. Furthermore, unlike the situation with oil- or gas-fueled
power plants, the cost of uranium fuel could increase greatly without having a serious impact on the cost of nuclear power. At $130
per kilogram, the cost of uranium contributes only one-third of a
penny to the cost of a kilowatt-hour, which is less than 5 percent of
the cost of electricity produced by a new light water reactor.3
Breeder reactors are costly to build and operate. About $100
billion (in 2007 dollars) has been spent worldwide on breeder reactor research and development and on demonstration breeder reactor projects. Yet none of these efforts has produced a reactor that is
economically competitive with a conventional light water reactor.
The capital costs per kilowatt of generating capacity of demonstration liquid sodium-cooled fast reactors have typically been more
than twice those of water cooled reactors of comparable capacity. Although it could be expected that once in production this cost
ratio would decline, today few, if any, experts argue that breeder reactor capital costs could be less than 25 percent higher than that of
similarly sized water cooled reactors.4
Detailed economic comparisons of light water and breeder reacBulletin of the Atomic ScientiStS | WWW.theBulletin.oRG may /j u n e 2 0 1 0
51
tors show that light water reactors operating on a “once-through”
fuel cycle (i.e., with spent fuel stored or disposed of in a deep geological repository) would be far less expensive than plutonium
breeder reactors under a wide range of assumptions.5
Fast neutron reactors have special safety problems. Plutonium breeder reactors require fast neutrons if they are to breed more plutonium
If the core of a breeder reactor heats
than they consume. They therefore canup to the point of collapse and suffers a
not use water as a coolant since neutrons
meltdown, the fuel can assume a more critical are slowed down dramatically by colliconfiguration and blow itself apart in a small
sions with the light nuclei of hydrogen in
nuclear explosion.
water. To date, the coolant that has been
used in all demonstration breeder reactors is liquid sodium.
Although sodium has some safety advantages, it also has serious drawbacks.
The most serious: It reacts violently with
water and burns if exposed to air. The
steam generators used in fast reactor systems, in which molten sodium and high-pressure water are separated by thin metal pipes,
have proven to be one of its most troublesome features. Any leak
results in a reaction that can rupture the tubes and lead to a major
sodium-water fire.
A large fraction of the liquid sodium-cooled reactors that have
been built have been shut down for long periods by such fires. A
notable exception is Russia’s BN-600 reactor, which has had a respectable operating record despite 14 sodium fires between 1980
and 1997. The reactor continued to operate throughout that time because its steam generators were located in separate bunkers to contain sodium-water fires and an extra steam generator was included
so the reactor could continue to operate while a fire-damaged steam
generator was repaired. Sodium leaks from pipes into the surrounding air also have resulted in serious fires. Japan’s Monju prototype
fast reactor experienced a major sodium-air fire in 1995. The reactor
may restart this year after 15 years of being off-line.
Sodium creates other problems as well. When it absorbs a neutron, it turns into sodium 24, a gamma-emitting radioactive isotope.
Thus, the sodium that cools the core becomes intensely radioactive.
To ensure that a steam generator fire does not disperse radioactive
sodium, reactor designers insert an intermediate sodium loop in
which heat generated from the reactor is transferred to non-radioactive sodium through a sodium-sodium heat exchanger. The nonradioactive sodium delivers heat to the steam generators that then
generate electricity. The extra sodium loops and associated pumps
contribute to the higher capital costs of breeder reactors.
Bulletin of the Atomic ScientiStS | WWW.theBulletin.oRG may /j u n e 2 0 1 0
52
Finally, unlike water cooled reactors that cease functioning if
they lose their coolant (a vital safety feature), fast neutron reactors
can actually become more reactive in cases where sodium coolant
is lost. Furthermore, if the core heats up to the point of collapse and
suffers a meltdown, the fuel can assume a more critical configuration and blow itself apart in a small nuclear explosion. Whether
such an explosion could release enough energy to rupture reactor
containment and cause a Chernobyl-scale release of radioactivity
into the environment is the subject of major concern and debate.
These fears convinced Germany to scrap its roughly $4 billion Kalkar demonstration fast breeder reactor and sell it for $3.3 million to
an amusement park operator.6
Sodium cooled breeder reactors have severe reliability
problems. The necessity of keeping air from coming into contact
with the sodium coolant makes refueling and repairing fast reactors
much more difficult and time-consuming than for water cooled reactors. The fuel has to be removed in an atmosphere free of oxygen,
the sodium drained, and the entire system flushed carefully to remove residual sodium without causing an explosion. Such headaches
have contributed to many fast reactors sitting idle a large fraction of
the time. France’s defunct Superphénix, the world’s only commercial-sized breeder reactor, generated on average less than 7 percent
of its capacity over its nominal operating lifetime. Japan’s Monju and
Britain’s Dounreay prototype fast reactors and the U.S. Enrico Fermi
1 demonstration breeder reactor had similar records. Russia’s BN600 has managed to maintain a respectable capacity factor (the percent of time it runs at full power), but only because of the willingness of its operators to keep it running despite multiple sodium fires.
Fast reactors and their fuel cycles pose serious proliferation risks. All reactors make plutonium in their fuel, but breeder
reactors require that this plutonium be separated from the ferociously radioactive fission products in spent fuel and reused. The
separation process, so-called reprocessing, also makes the plutonium more accessible to aspiring nuclear weapon makers.
This concern is not just theoretical. India justified its reprocessing program by citing an interest in breeder reactors, but in 1974, it
used its first batch of separated plutonium to carry out a “peaceful
nuclear explosion.” This led the United States to rethink its promotion of plutonium as the fuel of the future. France used its Phénix
prototype breeder reactor, starting in the late 1970s, to make weapon-grade plutonium. Decades of plutonium separation for planned
breeder reactor-based nuclear energy programs have produced
approximately 250 tons of stockpiled weapons-usable plutonium,
enough to make more than 30,000 nuclear bombs.7
The George W. Bush administration proposed a program to make
Bulletin of the Atomic ScientiStS | WWW.theBulletin.oRG may /j u n e 2 0 1 0
53
reprocessing more “proliferation resistant” by leaving some neptunium, americium, and curium mixed with the plutonium. Even if
this were done, however, the gamma radiation emitted by the mixture would still be one hundred times less than what the IAEA considers sufficient to protect against theft and thousands of times less
than the radiation protecting plutonium
when it is in spent fuel.
The persistence of breeder programs in
Future prospects for breeders. After
russia, India, and China is testimony to the
six decades and the expenditure of the
ability of their nuclear establishments to
equivalent of about $100 billion, the promtap into national treasuries despite the fact
ise of breeder reactors remains largely unthat breeders will not be able to compete
fulfilled. Britain, Germany, and the United
States have abandoned their breeder rewith light water reactors for the foreseeable
actor development programs. Despite the
future.
arguments by France’s nuclear conglomerate AREVA that fast reactors will ultimately fission the increasingly large stockpile of plutonium in France’s spent fuel,
the country has no operating fast reactors. Superphénix was shut
down in 1996 and Phénix in 2009. No replacement breeder reactor
is planned for at least a decade.
In Japan, there are plans for a new demonstration reactor by 2025
and commercialization of breeder reactors after 2050, but there is
reason to doubt these projections. The Japanese government is not
willing to kill its breeder program entirely, because, as in France,
the breeder is still the ultimate justification for shipping spent fuel
from individual reactors to a central reprocessing site. For decades,
however, Tokyo has been reducing funding for its breeder program
and shifting commercialization further and further into the future.
Russia and India are building demonstration breeder reactors,
and China is considering buying two such reactors from Russia.
The persistence of breeder programs in these countries is testimony to the ability of their nuclear establishments to tap into national
treasuries despite the fact that breeders will not be able to compete
with light water reactors for the foreseeable future. In India, the
breeder also is justified by its ability to greatly expand the rate of
India’s production of weapon-grade plutonium. Successful negotiation at the U.N. Conference on Disarmament of the Fissile Material
Cutoff Treaty, which would ban the production of fissile materials
for weapons, could foster a new debate in India about the value of
pursuing breeder reactors.
In the United States, during the Bush administration, fast reactors returned to the agenda reconfigured as “burner” reactors to
fission plutonium and other transuranic isotopes (e.g., americium,
neptunium, and curium) that are accumulating in U.S. spent fuel.
Bulletin of the Atomic ScientiStS | WWW.theBulletin.oRG may /j u n e 2 0 1 0
54
Theoretically, the fast neutrons in sodium cooled reactors would
indeed be more effective at fissioning this material than the slow
neutrons in water cooled reactors. But in 1996, a massive National
Academy of Sciences assessment commissioned by the Energy Department concluded that such an effort would have high costs and
marginal benefits and would take hundreds of years to greatly reduce the global inventory of transuranic isotopes. President Barack
Obama and the Democratic majority in Congress share this skepticism and have proposed a new research and development program
to investigate alternative strategies for managing U.S. spent fuel.
The breeder reactor dream is not dead, but it has receded far
into the future. In the 1970s, breeder advocates were predicting that
the world would have thousands of breeder reactors operating this
decade. Today, they are predicting commercialization by approximately 2050. In the meantime, the world has to deal with the hundreds of tons of separated weapons-usable plutonium that are the
legacy of the breeder dream and more being separated each year by
Britain, France, India, Japan, and Russia.
In 1956, U.S. Navy Admiral Hyman Rickover summarized his experience with a sodium cooled reactor that powered early U.S. nuclear submarines by saying 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.” More than 50 years later, this summary remains apt. <
Thomas B. Cochran is a senior scientist at the Natural Resources
Defense Council’s Nuclear Program. Harold A. Feiveson is a senior research scientist at Princeton University’s Program on Science
and Global Security. Zia Mian directs the Project for Peace and Security in South Asia at Princeton University’s Program on Science
and Global Security. M. V. Ramana is a visiting research scholar at
Princeton University’s Program in Science, Technology, and Environmental Policy and the Program on Science and Global Security.
Mycle Schneider is an independent energy consultant and produces
the “World Nuclear Industry Status Report.” Frank N. von Hippel
is professor of public and international affairs at Princeton University and co-chair of the International Panel on Fissile Materials.
nOTeS
1. Thomas B. Cochran, Harold A. Feiveson, Walt Patterson, Gennadi Pshakin,
M.V. Ramana, Mycle Schneider, Tatsujiro Suzuki, and Frank von Hippel, “Fast
Breeder Reactor Programs: History and Status,” International Panel on Fissile
Materials (IPFM) Research Report no. 8, February 2010. Available at http://www.
ipfmlibrary.org/rr08.pdf.
2. These projections were discussed in Mycle Schneider, Steve Thomas, Antony
Bulletin of the Atomic ScientiStS | WWW.theBulletin.oRG may /j u n e 2 0 1 0
55
Froggatt, and Doug Koplow, “2009 World Nuclear Industry Status Report,” Bulletin
of the Atomic Scientists, November/December 2009, vol. 65, no. 6, pp. 1–19.
3. For a full technical analysis of how much the cost of uranium factors into the
cost of electricity generated at a nuclear power plant, see Massachusetts Institute
of Technology, The Future of Nuclear Power, An MIT Interdisciplinary Study (Cambridge, MA: MIT Press, 2003), app. 5A.
4. This would be a capital cost difference on the order of $1,000 per kilowatt of
generating capacity compared to a conventional reactor. With a 10 percent capital
charge and a 90 percent average capacity factor (where the reactor operated 90
percent of the time), it would translate to a cost difference of about 1.3 cents per
kilowatt hour.
5. Matthew Bunn, Steve Fetter, John Holdren, and Bob van der Zwaan, “The Economics of Reprocessing Versus Direct Disposal of Spent Nuclear Fuel,” Nuclear
Technology , vol. 150 (June 2005), p. 209.
6. Klaus Janberg & Frank von Hippel, “Dry-Cask Storage: How Germany Led the
Way,” Bulletin of the Atomic Scientists, September/October 2009, vol. 65, no. 5, p. 29.
7. IPFM, Global Fissile Material Report 2009: The Path to Nuclear Disarmament.
Available at www.fissilematerials.org/ipfm/site_down/gfmr09.pdf.
Thomas B. Cochran, harold a. Feiveson, Zia mian, m. v. ramana, mycle Schneider &
Frank n. von hippel, “It’s time to give up on breeder reactors,” Bulletin of the Atomic
Scientists, may/june 2010, vol. 66, no. 3, pp. 50–56.
DOI: 10.2968/066003007
Copyright © 2010 Bulletin of the Atomic Scientists. All Rights Reserved.
Bulletin of the Atomic ScientiStS | WWW.theBulletin.oRG may /j u n e 2 0 1 0
56