Plutonium: How Nuclear Power’s Dream Fuel Became a Nightmare

© Springer Nature Singapore Pte Ltd. 2019

F. von Hippel et al.Plutoniumhttps://doi.org/10.1007/978-981-13-9901-5_1

1. Overview

Frank von Hippel¹  , Masafumi Takubo²   and Jungmin Kang³  

(1)

Program on Science and Global Security, Princeton, NJ, USA

(2)

3-93-9 Miyamoto-cho, Koshigaya, Saitama Prefecture, Japan

(3)

8340 Greensboro Dr. Unit 1025, McLean, VA 22102, USA

Frank von Hippel (Corresponding author)

Email: [email protected]

Masafumi Takubo

Email: [email protected]

Jungmin Kang

Email: [email protected]

One of the first tasks of the secret US World War II nuclear-weapon project was to design reactors to produce plutonium for bombs. This part of the project was headquartered at the University of Chicago, where, on 2 December 1942, a team led by Enrico Fermi and Leo Szilard, two European refugee physicists, created the first artificial steady fission chain reaction, which was sustained by neutrons traveling between lumps of uranium inside a pile of graphite.

After the operation of the pile confirmed their understanding of how a chain reaction could be achieved and controlled, the team worked with the DuPont company to design and build three hulking high-power plutonium-production reactors at the Hanford site on the Columbia River in remote eastern Washington state. These reactors produced the plutonium for the first test nuclear explosion in the southern New Mexico desert on 16 July 1945 and for the bomb that destroyed Nagasaki on 9 August 1945. After the war, 11 additional production reactors were built, and the 14 reactors together produced plutonium for the tens of thousands of additional nuclear weapons that the United States built during the Cold War.

In 1944, with the Hanford reactors about to go into operation, Fermi moved to Los Alamos, New Mexico, to work on the design of the plutonium bomb. In Chicago, Szilard and a few others in the reactor-design team started thinking about how to use nuclear energy to generate electrical power. They worried, however, that not enough high-grade uranium ore would be found to make fission energy into a significant energy source.¹ Chain-reacting uranium-235 constitutes only 0.7% of natural uranium. Virtually all of the rest is non-chain-reacting U-238.

1.1 Dreams of Plutonium Breeder Reactors

In the Hanford reactors, for every 10 atoms of U-235 consumed, about seven atoms of U-238 were being turned by neutron absorption into the artificial chain-reacting isotope plutonium-239. Neutrons released by the fissioning of the Pu-239 atoms could in turn convert more U-238 into Pu-239.

Szilard wondered whether more than one atom of plutonium could be produced from U-238 for every atom of plutonium fissioned. In that case, the resource base for nuclear power would become U-238 and it would be possible to produce about 100 times as much energy from the same amount of uranium. In fact, if the 3 grams of U-238 in an average ton² of crustal rock could be turned into Pu-239 and fissioned, the energy released would be 10 times greater than is released by burning a ton of coal.

Therefore, if what Szilard called the plutonium breeder reactor could be designed, civilization’s energy problems would be solved for millennia. This is why the breeder reactor has been called a dream machine.³

Szilard looked at the number of neutrons produced by the fission of plutonium as a function of the energy of the incoming neutrons and found that his idea could work for a fission chain reaction mediated by fast neutrons that still contain a significant fraction of the energy with which they were produced in the fission process.

In the Hanford reactors, neutrons were deliberately slowed down by collisions with the carbon nuclei of the graphite neutron moderator that makes up the body of the reactors. This is because only slow neutrons can sustain a chain reaction in natural uranium with its small fraction of chain-reacting U-235. If a reactor were fueled with a mixture of 15% or more plutonium in U-238, however, a chain reaction could be sustained with fast neutrons.⁴

But a reactor requires coolant to remove the fission heat from its core, and the water used as a coolant in most reactors today contains hydrogen that is very effective in slowing neutrons. The single protons that comprise the nuclei of hydrogen weigh almost exactly the same as the neutrons that mediate fission chain reactions. In the collision of a neutron with a proton—as in a collision between a fast and a stationary billiard ball of equal mass—the neutron can end up losing a large fraction or even all of its energy to the proton.

Szilard therefore searched for a coolant with heavier nuclei that neutrons would bounce off with much less energy loss—as a billiard ball would bounce off a cannonball. Ultimately, he settled on molten sodium.⁵ Its nuclei weigh 23 times more than hydrogen nuclei and are relatively nonabsorbing of neutrons. As a metal, it is also highly conductive of heat. The melting point of sodium is a relatively low 98 °C. If the reactor were kept above that temperature, the coolant would remain liquid even when the reactor was not operating.

The same considerations that were persuasive to Szilard resulted in liquid sodium metal being used as a coolant in almost all subsequent efforts to develop breeder reactors.⁶

1.2 Downsides of Breeders

It did occur to Szilard that there was a downside to the idea of powering the world with plutonium. In a 1947 speech on Atomic Energy, a Source of Power or a Source of Trouble, he spoke enthusiastically of the enormous energy resource that could be tapped using breeder reactors . But then, he added,

Unfortunately plutonium is not only an important atomic fuel, it is also the chief ingredient of atomic bombs. Can we afford to have atomic power unless we are safe from bombs? And can we be safe from bombs unless we can count on peace?⁷

Szilard’s invention attracted much more interest in the nuclear community than his warning, however, and a worldwide effort to develop plutonium breeder reactors began. Experimental Breeder Reactor I (EBR-I) went critical in 1951 at the US Atomic Energy Commission’s (AEC’s) National Reactor Testing Station (now the Idaho National Laboratory). It was followed by the United Kingdom’s Dounreay Fast Reactor (1959), France’s Rapsodie (1967), and the Soviet Union’s BOR-60 (1969).

A major technical challenge for sodium-cooled reactors is that sodium burns on contact with air or water. Very complex arrangements therefore are required to exclude air while refueling breeder reactors. Repairs inside of the reactors or their piping are time consuming because every trace of sodium has to be removed before the reactor or its piping can be opened up. And, if the thin metal barriers separating hot liquid sodium and water in one of the power plant’s steam generators leak, the resulting fire could destroy the steam generator.

Fast-neutron reactors also have another nuclear safety issue . In a water-cooled reactor, if the water overheats and boils, the steam bubbles make it less dense, its slowing-down effect on neutrons is reduced, a smaller fraction of them are captured by U-235, and the chain reaction, which is dependent on slow-neutron fissions of U-235, shuts down.

In contrast, in a plutonium-fueled breeder reactor, in which fast neutrons sustain the chain reaction, if the sodium boils and becomes less dense, the neutrons move faster, more neutrons are created per fission, and the power increases. The result could be a meltdown of the core into a still more reactive configuration resulting in a small nuclear explosion powerful enough to destroy the reactor containment and disperse the radioactive core into the atmosphere. The core of the very first fast-neutron reactor, EBR-I, had a partial meltdown. The first small breeder reactor operated by an electric utility, Fermi-1, 40 km from Detroit, also had a partial core-melt accident in 1966. That accident inspired both a book and a song titled We Almost Lost Detroit.⁸ During the nine years it was licensed to operate, Fermi-1 had so many problems that it produced the equivalent of less than one month of full-power output.⁹

1.3 Much More Uranium Found and Demand Growth Much Lower Than Projected

While the first generation of nuclear engineers involved in the development of civilian nuclear power plants worked on breeder reactors, the US Navy was in a hurry and developed reactors for submarine propulsion that were cooled and moderated by ordinary water. Such reactors are called light-water reactors (LWRs) to distinguish them from the heavy-water reactors that were developed by Canada.¹⁰ The first US nuclear power plant, Shippingport, on the Ohio River below Pittsburgh, Pennsylvania, was an LWR originally designed to power an aircraft carrier. The plant had an electric-power generating capacity of 60 megawatts electric (MWe) and began operating in 1957. It became the model for the 1,000-MWe-class LWRs that dominate nuclear power today.

By the 1970s, known resources of uranium had increased a thousandfold, and concerns about near-term uranium shortages had faded beyond any realistic planning horizon. In addition, after the 1979 partial meltdown of the core of the Three Mile Island Unit 2 power reactor in the United States, safety requirements were ramped up and both the capital and operating costs of nuclear-power reactors increased. This reduced the relative contribution of uranium to the cost of nuclear power. In 2018, uranium accounted for only a few percent of the cost of electricity from nuclear power plants.¹¹

Also, nuclear power’s demand for uranium fuel was much less than had been expected. As of the end of 2018, the global reactor fleet had a total generating capacity equivalent to about 400 1-gigawatt electric (1 GWe, or 1,000 megawatt) power reactors.

This fleet was very much smaller than the one that had been projected four decades earlier by the US Atomic Energy Commission and the International Atomic Energy Agency (IAEA), and its composition was significantly different.

In 1975, the IAEA had projected that global nuclear capacity in the year 2000 would be about 2,000 GWe, of which about 10% would be breeder reactors.¹² The IAEA did not project further into the future, but the US AEC projected in 1974 that the United States alone would have 2,300 GWe of nuclear capacity in 2010 and that about three-quarters of that capacity would be breeder reactors.¹³

As it turned out, as of the end of 2018, the United States had about 100 GWe of LWR capacity and no breeder reactors. Worldwide, the IAEA counted 454 operational power reactors with a total capacity of 402 GWe of which two were Russian sodium-cooled prototype breeder reactors. India had a prototype breeder reactor nearing completion, and China had a small 20 MWe experimental fast reactor, which produced power for the equivalent of an hour during its first five and a half years of operation.¹⁴

1.4 Reprocessing Spent Power-Reactor Fuel

In the 1960s and 1970s, with the global nuclear establishment expecting hundreds of breeder reactors to be built by the end of the century, several major industrial states launched programs to acquire plutonium for startup cores for the breeders. This was done by chemically reprocessing spent LWR fuel, which contains about 1% plutonium, using the technology developed by the United States during the Cold War to separate plutonium for nuclear bombs. Given a requirement of about 10 tons of plutonium to provide the first and second cores for a 1 GWe breeder reactor, 2,000 tons of separated plutonium would have been required to start up the 200 breeder reactors the IAEA projected for the year 2000. This is about 10 times as much as the quantity of plutonium separated for nuclear bombs during the Cold War .

France, Germany, Russia, the United Kingdom, and the United States all began constructing large civilian reprocessing plants, and Japan joined them in the 1990s. Only the French, Japanese, and UK plants were completed, however, and, as of the end of 2018, Japan’s plant had still not been licensed for regular operation.

1.5 A Wake-up Call from India’s Nuclear Test

In the United States, the initial impetus for reconsidering reprocessing was India’s first nuclear test explosion on 18 May 1974. The US Atoms for Peace program had supported India’s focus on nuclear power as a key to its modernization and had provided technical advice for its breeder and reprocessing programs.

India’s government insisted that its test, Smiling Buddha, was of a peaceful nuclear explosive that would be used for fracturing rock to release oil and other purposes—ideas that the US nuclear-weapon laboratories had promoted while they were fighting proposals for a ban on nuclear testing. Most US government experts concluded, however, that India had used a nominally civilian reprocessing effort to launch a nuclear-weapon program.

The United States also woke up to the fact that Brazil , South Korea , Pakistan , and Taiwan—all under military rule at the time—were negotiating to buy reprocessing plants. President Gerald Ford’s administration decided to block these purchases and, ultimately, none of the plants were built.

In 1977, Jimmy Carter became president as a critic of US promotion of a plutonium economy and launched a review of the rationale for the US breeder development program. He concluded—despite the opposition of the US nuclear-energy research and development (R&D) establishment and its congressional and foreign supporters—that breeders were both unnecessary and uneconomic. They were unnecessary because much more uranium had been discovered, including in the United States, and they were uneconomic because of the high cost and low reliability of liquid-sodium-cooled reactors when compared to water-cooled reactors.

The Carter administration therefore halted work on the Clinch River prototype breeder reactor in Tennessee and the licensing of a large civilian reprocessing plant that was nearing completion in Barnwell, South Carolina, next to the Department of Energy’s Savannah River Site, which had been established in the early 1950s as a second production site for weapons plutonium. Congress opposed the termination of the breeder project but ultimately canceled it after Carter left office because its costs kept climbing. The Ronald Reagan administration, which followed, was willing to allow commercial reprocessing to go forward, but without government subsidies. US nuclear utilities quickly concluded that, in the absence of a demand for separated plutonium, direct disposal of their spent fuel in a deep underground repository would be less costly than reprocessing. The Barnwell reprocessing plant was left uncompleted.

In Germany and Austria, anti-nuclear-energy movements, energized by the catastrophic Chernobyl reactor accident in 1986, besieged the construction site for a large German reprocessing plant near the village of Wackersdorf, Bavaria, near the border of Austria.¹⁵ Ultimately, in 1989, Germany’s nuclear utilities decided that it would be both less trouble and less costly to have their fuel reprocessed in France and the United Kingdom , which had both decided to use the expertise and infrastructure that they had developed in their military reprocessing programs to separate plutonium for domestic and foreign nuclear utilities.

Germany’s utilities continued to ship their spent fuel abroad to be reprocessed, even after the national breeder program collapsed in 1991. They had no alternative offsite destinations for their spent fuel. Most of the reprocessing contracts stipulated, however, that the plutonium and high-level radioactive waste from reprocessing would be returned to the owning country. A decade later, there were mass anti-nuclear protests again in Germany opposing the return of reprocessing wastes from France.¹⁶

In the Soviet Union, construction of a civilian reprocessing plant in a military plutonium-production center near Krasnoyarsk, Siberia, ended in the 1990s because of a lack of funds. A pilot reprocessing plant has been built there, however, and reprocessing of power-reactor spent fuel continued at a small converted military reprocessing plant in Ozersk in the Urals.

As of 2019, the Russian and Indian nuclear establishments were continuing to push their breeder-reactor development programs slowly forward. In China, the government-owned China National Nuclear Corporation (CNNC) had built small experimental civilian reprocessing and breeder facilities. Both were operational failures. Nevertheless, CNNC had begun construction on intermediate-scale plants.

1.6 Plutonium Fuel for Light-Water Reactors

In 1998, France shut down Superphénix, the highest-power breeder reactor ever built (1.2 GWe). Due to sodium and air leaks and other problems, the reactor had produced in 12 years the equivalent of only one-third of a year of full-power output.¹⁷

And what about the accumulating plutonium that had been separated for breeders that had not been built? France’s nuclear establishment argued that fast-neutron reactors would be built eventually and that reprocessing should be sustained in the meantime, by recycling excess separated plutonium in conventional LWR fuel.¹⁸ France then began fabricating mixed-oxide (MOX) fuel from a mixture of plutonium and depleted uranium oxides. (Depleted uranium is the waste product from uranium enrichment.) In 1995, following joint operation with Belgonucléaire of two small MOX-fuel plants in Belgium and France, France started up the large MELOX MOX-fuel fabrication plant at its Marcoule site. Use of MOX fuel in LWRs reduced France’s requirements for low-enriched uranium (LEU) by about 10%.

Japan’s nuclear-energy establishment decided to take a similar approach—initially by having its spent fuel reprocessed in France and the United Kingdom, where the separated plutonium would be fabricated into MOX fuel before being sent back to Japan. The use of the resulting MOX fuel in Japan was delayed by a decade, however, by public opposition aroused in part by a scandal in 2000 over falsification of quality-control data for the first batch of MOX fuel made in the United Kingdom. The 2011 Fukushima accident delayed the program further.

In any case, the economics of MOX fuel are terrible. In France in 2000, a rigorous assessment commissioned by Socialist Prime Minister Lionel Jospin concluded that, including the cost of reprocessing, the production of MOX fuel would cost five times as much as the LEU fuel that otherwise would have been used.¹⁹

There were benefits for France and the United Kingdom, however, in the form of foreign exchange from reprocessing. Utilities in Japan, Germany , Switzerland, Belgium , and the Netherlands contracted with France and the United Kingdom to have their spent fuel reprocessed and to have the recovered plutonium fabricated into MOX fuel. In the United Kingdom, the reprocessing of foreign LWR spent fuel was the entire rationale for the construction of the Thermal Oxide Reprocessing Plant (THORP) and the Sellafield MOX Plant. THORP reprocessed the spent fuel from the United Kingdom’s advanced gas-cooled reactors as well, but the resulting separated plutonium simply went into storage.

In 2000, as part of a larger decision to phase out Germany’s nuclear-power program, however, Germany decided not to renew its reprocessing contracts.²⁰ Japan also did not renew its reprocessing contracts with France or the United Kingdom, but for a very different reason: it had launched construction on its own large reprocessing plant. In the United Kingdom, the main remaining customer for reprocessing was Électricité de France (EDF), France’s nuclear utility, which had taken over ownership of the UK advanced gas-cooled reactors and the UK’s single LWR. EDF also declined to renew.

With both its foreign and domestic customers having decided not to renew their reprocessing contracts, the UK government had no alternative but to end its reprocessing program. It established a Nuclear Decommissioning Authority to complete the existing contracts and then carry out the cleanup of its plutonium and reprocessing site at Sellafield. In 2018, the cost of that cleanup was estimated at £91 billion ($120 billion).²¹ THORP shut down at the end of 2018.²² As of 2018, operations at an older plant reprocessing the spent fuel from the United Kingdom’s shut-down first-generation Magnox reactors were expected to end in 2020.²³

In France, despite the loss of foreign reprocessing customers, the government forced EDF to renew its reprocessing contract to support the government-owned nuclear fuel services and reactor construction company. That company, Areva, hemorrhaged funds in its reactor-construction and uranium-mining operations. In January 2018, it was reorganized as a smaller company, Orano, focused on uranium enrichment and spent-fuel reprocessing.

Globally, civilian plutonium separation vastly exceeded plutonium use in fuel. Consequently, despite the end of the Cold War, the global stockpile of separated plutonium continued to grow, with the civilian stockpile reaching about 300 tons in 2018 (Fig. 1.1).

Fig. 1.1

Global stock of separated plutonium. The global stock of weapons plutonium plateaued at the end of the Cold War, and most of it has become excess as a result of dramatic reductions in the numbers of US and Russian operational nuclear warheads. The global stock of civilian separated plutonium continued to grow, however, because of continued reprocessing of spent fuel despite the failure of breeder commercialization and the limited extent of plutonium recycle in conventional power reactors. Assuming 8 kg of plutonium for a warhead (the IAEA’s metric), current stocks of about 300 tons of civilian plutonium are sufficient to make more than 35,000 Nagasaki-type warheads. (IPFM, updated by authors)²⁴

1.7 Reprocessing for Radioactive-Waste Management?

With breeder-reactor