Skip to main content

Essay Contest: Physics in 2116

Accompanying Frank Wilczek’s article about what physics will look like in 100 years is an opportunity for our readers to submit their own predictions for the chance to win $7500. Find out about our Physics in 2116 contest online.

Why did the US abandon a lead in reactor design?

A questionable reshaping of reactor research 45 years ago has had long-term consequences.

Sometime in the late 1960s, a great shakeup occurred in nuclear reactor research. As a young employee of the reactor division at the Los Alamos Scientific Laboratory at that time, I was shocked and confused when the division was suddenly dissolved. Now that we are again considering alternatives to light-water reactors, several narratives have sprung up to explain why these alternatives were abandoned.

Recently, I decided to research that decision using publicly available sources. What I found was remarkable. The key player was Milton Shaw, who directed the Atomic Energy Commission’s (AEC) Reactor Development and Testing Division (RDTD) at that time. Shaw refocused the US civil nuclear program toward a single goal of the liquid-metal fast breeder reactor, making a number of strategic mistakes that have had long-term safety consequences for the industry.

Shaw was a protégé of Admiral Hyman Rickover, known as the father of the US nuclear navy. Rickover and his team successfully developed nuclear reactors for submarines and then aircraft carriers, releasing them from the need for fossil fuels as the main source of propulsion. The first nuclear submarine, USS Nautilus, was authorized by Congress in July 1951 and was launched January 1954, two and a half years later. The aircraft carrier USS Enterprise was authorized in 1954 and commissioned in 1958, less than four years later.

The reactor pressure vessel of the Shippingport Atomic Power Station during construction in 1956. The plant was commissioned two years later.

The reactor pressure vessel of the Shippingport Atomic Power Station during construction in 1956. The plant was commissioned two years later.

In December 1953 US President Dwight D. Eisenhower focused on peaceful uses of nuclear power in his “Atoms for Peace” speech at the United Nations in New York City. The fastest way to implement his vision was for Rickover’s team to adapt the navy’s USS Enterprise reactor design for civilian use. Within months, ground was broken for the first nuclear electrical generating plant at Shippingport, Pennsylvania. Less than four years later, the plant was operational. Shaw led the team.

These were impressive engineering and patriotic successes at a time when the US felt the Cold War pressure of competing with the Soviet Union. Rickover and his team were considered effective, although there were stories about bullying and dictatorial management practices.

Extending civilian nuclear power

The success at Shippingport, and other civilian reactors, raised Shaw’s visibility. In 1964 he was made director of the RDTD. Commercial firms were getting into the reactor business, extending the navy’s pressurized light-water reactor (PWR) design to larger sizes. However, nearly all the main reactor research was being conducted at the US national laboratories.

Argonne and Oak Ridge national labs were involved with the design of the first submarine reactor, which was built by Westinghouse Bettis Laboratory (now owned by Bechtel) and tested at the Idaho National Reactor Testing Station (now Idaho National Laboratory). Although the first reactors had been carbon-moderated and gas-cooled, water served both purposes in the naval reactors. The water also shielded the crew from radiation.

As more commercial reactors were built for power plants, their size and operating temperatures were increased. To serve their purpose, they would need to be sited close to populated areas. Though an earlier safety standard required that reactors be kept in desolate areas away from large population centers in case of an accident, reactors would be uncompetitive with other forms of power production if they were sited too far from their energy market.

Safety studies were duly carried out at the national labs on containment vessels as an alternative safety mechanism. The most damaging accident thought to be possible (“maximum credible accident”) was losing reactor water coolant when a pipe broke. Although losing the moderator would end the fission reactions, the heat generated by fission products might melt the core. A facility was planned at Idaho to test such an accident.

It was not uncommon at that time for engineers to test equipment to failure. For example, nuclear-powered rocket engines, which also being developed at the time, were tested at the Nevada Test Site. In 1965, as part of this program, the Kiwi-TNT test simulated the maximum credible accident for that reactor: a rapid nuclear excursion, such as might happen if the reactor powering a rocket fell into the ocean.

The national laboratories were also investigating a wide variety of reactor designs, from liquid-metal-cooled reactors through high-temperature gas-cooled reactors to molten salt reactors. The reactors might be fueled by uranium, plutonium, or thorium. Some of them, called breeders, made more fuel than they used. Breeders would make the most of what were then believed to be very limited uranium reserves by creating plutonium as the uranium was burned. Shaw’s Reactor Development and Testing Division funded both safety studies and exploratory reactor designs.

The pivot point

By the late 1960s, two breeder reactors had been operating in the US: An experimental breeder reactor (EBR-I) built in Idaho in 1949 to prove the principle of breeding and the Enrico Fermi breeder reactor near Detroit, Michigan, for which construction started in 1963. Both used liquid sodium as coolant, which has ideal nuclear properties for working with fast neutrons—useful for efficient breeding of plutonium. Sodium is a solid but soft metal at room temperature and reacts strongly with water. This, as researchers discovered, makes it difficult to work with: EBR-1 suffered a partial meltdown in 1955, as did Fermi in 1966, both from blocked coolant channels.

Starting in about 1968, the resources of Shaw’s Reactor Development and Testing Division were turned solely towards the development of a liquid metal fast breeder reactor (LMFBR), and all other projects were abandoned. Why?

The answer to that question would require more detailed research in the AEC archives. Open internet sources indicate that Shaw probably had the approval of AEC Chairman Glenn Seaborg and most of the other commissioners. The AEC, born out of the wartime Manhattan Project as the civilian agency overseeing all uses of nuclear power, was accustomed to acting in secrecy. There is some indication that the deliberations surrounding this decision were more secretive than others.

If RDTD focused narrowly on a single objective, this practice was consistent with prior naval reactor development. For one example, the second nuclear submarine, USS Seawolf, was originally fitted with a liquid-sodium reactor, which was replaced after a few years with the standard PWR used for nuclear submarines.

Postcard celebrating the launch on 21 July 1955 of USS Seawolf (SSN-575). The submarine’s liquid metal cooled reactor proved difficult to maintain. The submarine had a crew of 101 and was equipped with six 21-inch torpedo tubes.

Postcard celebrating the launch on 21 July 1955 of USS Seawolf (SSN-575). The submarine had a crew of 101 and was equipped with six 21-inch torpedo tubes.

At the time, with rising energy demands and predictions of uranium deposits running out, a breeder reactor program seemed the only logical choice if nuclear energy was to contribute significantly to US energy needs. Shippingport and the naval reactors had operated without any serious problems for a decade, which could be taken as evidence that safety issues had been mastered. Rickover was proud of his safety program, as indicated in his testimony to Congress in 1979. The result was that the national laboratories’ safety research was suddenly cut back and later ended, because it was believed unnecessary, or acceptable to delegate to industry. Research into alternative reactor designs at the labs also ended.

Shaw seems not to have consulted the senior staff of the national laboratories on the decision in any significant way. Alvin Weinberg, director of Oak Ridge, was reported to have been livid. The reactor division at Los Alamos was shut down. Work at Idaho was redirected multiple times. A program manager for the LMFBR was installed at Argonne who reported directly to Shaw.

In response to Shaw’s initiatives in 1967, Albert V. Crewe, then director of Argonne National Laboratory, noted that Argonne’s purpose was “not to build submarines but to produce knowledge.”

Shaw’s application of Rickover’s narrow focus and command-hierarchal structure marked a break with past management practices of the national laboratories. Until then, reactor programs at the national laboratories had been developed through consultation between the AEC and the laboratories, and allowed for exploratory projects. This broad view came from the Manhattan Project, in which multiple pathways toward the goal of an atomic weapon were investigated.


Narrowing the government’s reactor program to sole emphasis on the LMFBR probably had more influence on the shape of today’s nuclear industry in the US—and to its opposition—than any other single decision after the Manhattan Project. Shaw’s program failed in its primary objective, to build a prototype LMFBR at Clinch River, Tennessee. It also left vulnerabilities in the light-water reactor designs that would be used in power plants around the world. Few alternative reactor designs, such as gas-cooled reactors, were built into US power plants, and those few were commercially unsuccessful.

The LMFBR program failed for many reasons. It moved more slowly and cost more than was expected. Concerns about proliferation rose during the 1970s. Additional reserves of uranium were found, making breeders less attractive. Congress’s Joint Committee on Atomic Energy—dissolved when AEC was split into the Energy Research and Development Administration (ERDA) and the Nuclear Regulatory Commission (NRC)—was no longer available to provide single-minded financial support. In contrast, France, Russia, China, and Japan all operated commercial-scale liquid-metal-cooled fast reactors.

The AEC’s dual role as both regulator and developer of nuclear energy was often questioned and criticized. The cut in safety research was the final straw that led to the 1975 splitting of the AEC into ERDA and NRC. Two years later, ERDA was combined with the Federal Energy Administration to form the DOE. Like the regulatory arm of the AEC, the NRC has no funding for safety research but over the years DOE has sporadically addressed some reactor safety concerns, as do the reactor manufacturers.

The narrowed focus of the US reactor program left dangerous gaps. Both the Three-Mile Island (1979) and Fukushima (2011) reactor meltdowns were due to loss of coolant, the type of accident on which Shaw shut down research. Modifications now being proposed might have been introduced years ago, in time to prevent those accidents, if that research had continued.

In fact, because that research was curtailed, some of the scientists in these divisions took their safety concerns and expertise to people and organizations that would listen to them. The Union of Concerned Scientists, formed in 1969, and other groups benefitted from their knowledge transfer. Such groups would have emerged in any case, but the alienation of scientists working on reactor safety and the real problems they addressed strengthened their resolve and mobilization.

Development is restarting on reactor designs that were mothballed by Shaw’s decision nearly 50 years ago. Thorium, liquid-salt-fueled, and liquid-metal-cooled reactors are being considered by the national labs, some universities, and startup firms. So too are more innovative, safer versions of light water reactors. None of those designs can be expected to come into use for a decade or more. Corporate research will be very focused, and we cannot expect that safety issues will be a primary consideration. Contrary to claims by some promoters of the new designs, experiments conducted through the 1960s are barely proof of concept. Regulations and available technology have changed too much for the earlier data to be reliable.

Not a positive story

Shaw’s redirection of the American nuclear program is little mentioned in official histories on the web. Its effect on the breakup of the AEC is partially recognized on the NRC’s history page, and not at all in a DOE history of ERDA. Undoubtedly relevant documents exist in AEC and national laboratory archives.

The story is not a positive one for the AEC. Although most biographies recognize that Admiral Rickover’s management of his programs was controversial, he is still revered in the US nuclear navy and beyond. In the civilian industry, a standard term for nuclear power reactors collectively is “nuclear fleet,” recalling their naval origins. There is little motivation to revisit that history, but understanding the strategic mistakes can help us chart a more productive way forward.

Shaw and the AEC failed to recognize the differences between developing naval reactors and developing the LMFBR: the comparatively more technical unknowns requiring investigation, the culture and history of the national laboratories in contrast to the captive naval reactor labs, and the need to avoid perceptions or realities of conflicts of interest. They eliminated alternative paths that might have provided better commercial reactors or developed safety fixes. Many of the negative consequences have endured for forty years and more. Perhaps looking back at what went wrong can help to repair the damage.

Cheryl Rofer worked as a chemist at Los Alamos National Laboratory. Now retired, she contributes to the online forum Nuclear Diner.

Submit comment
Comment moderation successfully completed
ec53185b92e297f26a707ada48e7a20c ptol.magazine_postzxybnytfddd