The Empty Vault
David Lilienthal was a lawyer from Illinois who had spent the New Deal years running the Tennessee Valley Authority. He had built dams. He had electrified rural Tennessee. He believed, with the particular intensity of a Midwestern progressive of his generation, that government could do great things if it were organized around the public interest rather than private gain. When Harry Truman asked him in October 1946 to chair the newly created Atomic Energy Commission, Lilienthal accepted reluctantly but with a sense of historical mission. The atom, he believed, could be turned from weapon to servant.
On the morning of April 3, 1947, Lilienthal drove to the White House to brief the President on what he had inherited.
The meeting was extraordinary in its frankness. The AEC had been in legal existence for only three months. Its commissioners had been confirmed by the Senate after a bruising fight in which Senator Kenneth McKellar of Tennessee had tried, and failed, to destroy Lilienthal’s reputation. The agency had taken physical custody of the wartime nuclear enterprise on January 1: Hanford, Oak Ridge, Los Alamos, the scattered laboratories and production plants that General Groves had overseen as the Manhattan Engineer District. Lilienthal had spent three months learning what he had.
What he had was less than the public believed. He told Truman — a few associates present, the information so sensitive that no one took notes — that the nation’s nuclear weapons stockpile was essentially nonexistent. The bombs assembled at Los Alamos were few in number. Production of fissile material had slowed to a trickle since the end of the war. Hanford’s wartime reactors were aging, and no one had yet worked out how to replace them. The nation had the scientific knowledge to build atomic weapons. It did not, as a practical matter, possess them in quantities that would matter in a war.
Truman listened without expression. Lilienthal later wrote that the President’s face darkened as the briefing continued. Churchill, at Fulton the previous year, had spoken of an iron curtain descending across Europe and had credited the American atomic monopoly with restraining Soviet aggression. That monopoly, Lilienthal was telling Truman, was a bluff.
Whatever the civilian possibilities of the atom — whatever power reactors might someday do, whatever medical isotopes might cure, whatever the visionaries of the Met Lab had dreamed of for the postwar world — the immediate task of the Atomic Energy Commission was now obvious. It had to produce plutonium. It had to produce a great deal of plutonium, and it had to do so quickly.
The institutional consequences of that morning’s briefing would shape American nuclear technology for the next three decades.
The Oppenheimer Committee
In July 1947, J. Robert Oppenheimer convened the first meeting of the AEC’s General Advisory Committee in a three-story conference room in the New War Department Building in Washington. He had been appointed chairman of the committee — a body of nine distinguished scientists charged with advising the Commission on scientific and technical matters — and he took the role seriously. The committee included James Conant, Enrico Fermi, Isidor Rabi, Lee DuBridge, and Glenn Seaborg, among others. It was, collectively, the most eminent scientific body that had ever advised the American government on anything.
Oppenheimer opened the meeting by asking Lilienthal to brief the committee on the weapons stockpile. Lilienthal, mindful that the information was now the most closely held secret in the government, insisted that military representatives be present and that the information be conveyed orally, with accuracy “within plus or minus twenty percent.” Robert Bacher, recently returned from Los Alamos, recited the numbers while the committee members closed their notebooks and listened. It was, as one participant later recalled, a dramatic moment.
The numbers confirmed what Lilienthal had told Truman. The stockpile was thin.
Over the following days, as Oppenheimer led the committee through a review of AEC priorities, the question emerged in concrete form: given limited resources, what should the Commission do? Every activity could not be given top priority. Weapons production required reactors, chemical separation plants, fabrication facilities, all of which competed for the same engineering talent and the same construction budgets. Research reactors competed with production reactors. Civilian power reactors, if they were ever to be built, would compete with both.
Oppenheimer, on the afternoon of the second day, began to think out loud. He understood perfectly the value of weapons — he had built the first ones, after all — but he could not, as a matter of conscience, give reactors a second priority. The committee had spent a year the previous spring, under his chairmanship, drafting the Acheson-Lilienthal report on international control of atomic energy. That report had foreseen a world in which atomic power would be taken out of the hands of national military establishments and placed under international control, so that its civilian applications could flourish without the shadow of weapons production. The report had failed politically — Baruch had revised it into a proposal the Soviet Union would not accept, and the Soviet Union had rejected it — but its underlying vision survived in Oppenheimer’s thinking. Reactors for power generation represented the most important peaceful application of atomic energy. Perhaps with top priority, he said, it might be possible to obtain useful power from a reactor within a year or two.
The sentence hung in the room. Everyone present knew it was not true. No reactor capable of generating useful civilian power existed in 1947, or could be built in one or two years. Oppenheimer knew it. Seaborg, who had spent the war on the production reactor problem, knew it. Fermi knew it.
But the sentence captured something Oppenheimer genuinely believed: that without the public promise of civilian power, atomic energy would be permanently enslaved to the weapons program. If the AEC was to be anything other than a production arm of the Department of Defense, it had to develop a civilian purpose. That civilian purpose, whatever the technical obstacles, had to be power.
The General Advisory Committee would spend the next several years struggling with this tension. Lilienthal himself was soon uncomfortable with the enthusiasm that public statements about nuclear power were generating. In July 1947, Oppenheimer proposed that the Commission issue a statement giving “a realistic evaluation of atomic power” — specifically, one that would counteract the growing misconception that economic nuclear power was imminent. Lee DuBridge warned that opinion was growing among scientists that there was no valid reason for the absence of practical nuclear power other than the Commission’s failure to act. Lilienthal considered it, and drew back. A realistic statement would deflate overoptimism, yes, but it would also hand ammunition to those who wanted to return atomic energy to military control. The civilian AEC, still in its first year, might not survive such a debate. The question was tabled.
This pattern — the private acknowledgment that civilian nuclear power was years away, coupled with public rhetoric that implied it was imminent — would recur throughout the AEC’s history. It was, perhaps, the original sin of the civilian program. It taught the Commission to promise more than it could deliver, and it taught the Congress and the public to expect more than was technically possible.
The Captain Arrives
Hyman George Rickover was a short, thin, wiry man with a high voice and an acid tongue. He had graduated from the Naval Academy in 1922, the son of a Polish-Jewish tailor from Chicago, and had spent the interwar years doing what most engineering duty officers did — serving on ships, earning a graduate degree in electrical engineering at Columbia, working in obscure technical billets in Washington. During the war he had commanded the electrical section of the Bureau of Ships. He was forty-seven years old when he arrived at Oak Ridge in June 1946.
Rickover was not the kind of officer who rose naturally through the Navy’s ranks. He was abrasive, contemptuous of hierarchy when hierarchy impeded technical work, willing to make enemies and disinclined to repair the damage afterward. Senior officers found him insubordinate. Junior officers found him tyrannical. What distinguished him from the many other competent engineers in the Bureau of Ships, however, was a quality that was rare in any bureaucracy: he was willing to accept full personal responsibility for the work of his subordinates, and he was willing to demand, from them, a standard of technical excellence that most organizations did not think possible. During the war, his electrical section had been notable for its obsession with inspection, with follow-up, with catching defects before they reached the fleet. He had walked through battle-damaged ships himself, in the yards, to see how his equipment had performed under fire.
In 1946, Admiral Earle Mills, the chief of the Bureau of Ships, sent Rickover to Oak Ridge with five other officers and three civilian engineers to learn what nuclear reactors were. The trip had been controversial within the Navy. Many senior officers thought Rickover was the wrong man to send. Mills sent him anyway. His only concession to the skeptics was that Rickover would not be in charge of the group; the officers would report to an Army colonel, and the civilians would be assigned to the Daniels reactor project, a proposed graphite-moderated high-temperature gas-cooled reactor that Farrington Daniels of Wisconsin was promoting.
What happened at Oak Ridge over the following year can be told briefly but is, in its implications, one of the most consequential episodes in the history of American technology.
Rickover, officially a student, made himself a center of technical learning. He attended lectures. He read reports. He asked questions of everyone willing to answer them. He visited the production reactors at Hanford. He visited Argonne. He met Alvin Weinberg, who was by then the research director at what had become Oak Ridge National Laboratory; Weinberg and Rickover would form a working relationship that was professionally productive even when personally strained. Rickover talked with Eugene Wigner about reactor theory. He talked with Seaborg’s former colleagues about plutonium chemistry. By the end of the year, he probably understood the state of American nuclear engineering as well as any individual in the country.
What he concluded was this: the fluid-fuel reactors that Wigner and Weinberg were thinking about — the thorium breeder of MUC-EPW-134 — were not yet reactors. They were concepts. The materials problems had not been solved. The chemistry had not been worked out at engineering scale. The corrosion behavior of high-temperature molten salts or aqueous uranium solutions in intense radiation fields was unknown. Everything about the fluid-fuel approach was harder to predict than everything about solid-fueled reactors. Solid fuel, pressurized water as both moderator and coolant, uranium metal clad in some corrosion-resistant jacket — this was a reactor you could calculate, test, and build.
For a submarine, calculable was essential. A submarine reactor had to run submerged, for months at a time, without maintenance, in spaces too small for extensive auxiliary equipment, with operators who could not be expert reactor physicists. It had to be reliable the way a ship’s boiler was reliable. Rickover looked at the fluid-fuel concepts and saw a research program. He looked at the pressurized-water concept and saw an engineering problem he could solve.
In 1947, Rickover returned to Washington with his conclusions. He would build a pressurized-water reactor for a submarine. He wanted Westinghouse as the prime contractor; he did not want General Electric, which was pursuing its own intermediate-spectrum breeder at Knolls and which Rickover suspected of trying to hijack the Navy program to subsidize its civilian ambitions. He wanted the work concentrated in a naval reactor branch that would report both to the Navy’s Bureau of Ships and to the AEC’s Division of Reactor Development, giving him institutional leverage in both organizations. He wanted personnel of his own choosing, trained to his own standards, dedicated to this single project.
He got all of it. By 1948, the Naval Reactors Branch existed. Westinghouse had built a new laboratory at Bettis, near Pittsburgh, staffed by engineers Rickover had personally recruited. An identical submarine reactor prototype was under construction at the Idaho desert testing station — Rickover’s insistence that a full land-based prototype be built before any seagoing reactor, a principle he would enforce for the rest of his career. The USS *Nautilus* would follow, achieving criticality in 1953 and sailing under the polar ice cap in August 1958.
What no one quite understood at the time, including Rickover himself, was that this was not merely the birth of the nuclear Navy. It was the template that would be imposed, by default, on the civilian nuclear industry.
The Young Veteran
On a November night in 1944, a Liberator bomber was returning to England from a daylight mission over Holland. Its pilot was a twenty-four-year-old American named William Liscum Borden, a law student who had interrupted his studies at Yale for the war. The cockpit was dark, the aircraft on autopilot, the crew tired. Then suddenly the sky above the plane lit up in crimson. Borden looked up and saw, passing above his aircraft, a V-2 rocket headed for London. It moved so fast that his Liberator, which had seemed impressive in its own right, appeared in comparison to be motionless — “a primitive horse-drawn buggy,” Borden would later write. The V-2 had left its launch site in Holland some four minutes before overtaking him. It struck London before his navigator finished making the entry in the flight log.
What Borden took from that moment was a conviction that would organize the rest of his working life. Rockets of the V-2 type would soon carry atomic warheads — the atomic bomb was not yet operational in November 1944, but its existence was no secret to people with security clearances. Rockets carrying atomic warheads would, sooner rather than later, reach across the oceans. The geographic isolation that had protected the Western Hemisphere through two world wars would not survive another decade. And if the next war would be fought with intercontinental atomic rockets delivered without warning, then everything about American strategic policy needed to change. The old pattern — a small standing military, an industrial base that mobilized slowly after hostilities began, a population that believed itself protected by oceans — would leave America defenseless. The next war would be over before the war economy could begin.
After V-E Day, Borden returned to Yale Law School and, in the evenings, wrote a book. It was published in 1946 under the title *There Will Be No Time: The Revolution in Strategy*. He was twenty-six years old when it appeared. It was the book of a very young man arguing that the great strategic question of the twentieth century had been settled on a November night over the English Channel: there could be no more mobilization strategies, because there would be no time to mobilize. Peacetime was the only time to build a military arsenal capable of winning the next war. Once the rockets started flying, the war would be decided by whatever stockpiles and forces had been assembled in advance.
The book laid out the argument in sequential chapters, each removing an excuse. The United Nations, Borden wrote, could not prevent war; it was “a scrap of paper as soon as a strong power elects to treat it as such.” International anarchy was a permanent condition that only a global monopoly on force could overcome, and the United States could not rely on such a monopoly arriving in time. Rockets had made naval blockade and strategic bombardment obsolete as the organizing principles of strategy, because both required time that the atomic age no longer allowed. The key chapter, “The Pattern of Atomic Warfare,” argued that an aggressor in the next war would strike first at military targets — airfields, missile sites, stockpile depots — rather than cities, because destroying American retaliatory capacity in the first minutes would win the war before civilian morale was even tested. Therefore the only sensible American strategy was to maintain, in peacetime, forces and stockpiles so large and dispersed that no first strike could eliminate them. Mobilization after war began was a nineteenth-century idea. The atomic age had abolished it.
The book sold modestly. It received respectful reviews in a few defense journals. In 1946, Borden was a law student with an unusual hobby, not a strategic theorist whom anyone in Washington needed to heed. He graduated from Yale Law in 1947 and began looking for work.
The opportunity came through a connection that would determine the direction of Borden’s career. Brien McMahon was a Democratic senator from Connecticut, a former Justice Department lawyer, newly returned to politics after the war. In 1946 McMahon had authored the Atomic Energy Act — the law that created the civilian AEC, placed atomic energy under congressional oversight, and established the Joint Committee on Atomic Energy as the body to which the AEC would answer. McMahon had chaired the Joint Committee at its founding in 1946, lost the chairmanship when the Republicans took the Senate in 1946, and regained it when the Democrats returned to the majority in 1949. He was forty-five years old in 1949, energetic, politically ambitious, and committed to the proposition that atomic weapons were the future of American military power. He was thinking about running for president.
McMahon read Borden’s book, and in 1949 hired him to serve as the Joint Committee’s executive director — the senior staff position overseeing the congressional oversight of American nuclear policy. Borden was twenty-eight years old when he took the job. He would hold it for four years, through the defining period of the American weapons complex. The thesis of his 1946 book — peacetime stockpile expansion as the only rational response to the atomic age — would, during those four years, be implemented as national policy on a scale and at a pace that no one in 1946 had imagined possible.
McMahon’s death from cancer in July 1952 would not stop the program Borden had helped to build. By then the expansion was so far along, and so institutionally entrenched, that no individual’s death could reverse it. What Borden had done in those four years, working through McMahon’s committee, was to ensure that when the Soviets tested their first atomic weapon in August 1949 and when North Korea invaded the South in June 1950, the congressional machinery was already in place to translate those shocks into concrete production programs. The Arsenal, as it rose, rose under pressure that Borden had done more than any other single staff-level figure to create.
The Detroit Initiative
In October 1950, while Rickover’s prototypes were under construction and while Weinberg at Oak Ridge was pursuing the aqueous homogeneous reactor experiments that he believed would lead eventually to a fluid-fuel thorium breeder, Walker Cisler of Detroit Edison and a team from Dow Chemical began a joint study of civilian nuclear power. They had AEC clearance to review classified reactor data. They spent a year doing exactly that — visiting every major nuclear installation in the country, reading the reports, interviewing the physicists.
Their conclusions have been described in the companion essay on Cisler that accompanies this book. The study group’s first preference was the liquid-metal-fueled thermal breeder using uranium-233. Their second was a modified heavy-water reactor. Their third was the fast plutonium breeder based on EBR-I. On November 8, 1951, at the McMath-Hulbert Observatory on Lake Angelus in Oakland County, Michigan, the technical committee drafted a statement that a high-temperature breeder using mobile fuel with fluid reprocessing represented the best approach.
One detail about their first choice deserves emphasis, because it would matter for what followed. The liquid-metal-fueled reactor Cisler’s engineers preferred was not a hypothetical. Brookhaven National Laboratory had been working on exactly this technology — uranium dissolved in molten bismuth, with a thorium blanket as a slurry of bismuth and thorium bismuthide — since 1947. By 1951, when Cisler’s engineers picked it, the Brookhaven Liquid Metal Fuel Reactor project was an organized program with designs and components in development. Cisler’s choice was not visionary speculation. It was a bet on a program the AEC was already funding, through a national laboratory, alongside Oak Ridge’s fluoride-salt work and several other threads of active fluid-fuel research.
What matters for the present narrative is what the Cisler group’s work represented institutionally. Cisler was not a government scientist. He was not a national laboratory director. He was the president of the nation’s third-largest electric utility, a man who had personally organized power infrastructure across war-ravaged Europe, and he commanded the attention of industry in a way that Weinberg did not. When Cisler said that the thorium breeder was the right technology, he was speaking not as an advocate but as a customer — as one of the men who would actually have to buy and operate whatever the AEC managed to develop. His judgment carried weight because it was commercially disinterested. He wanted a reactor that would generate power economically. He had concluded, after a year of careful study, that the fluid-fuel thorium breeder would do that better than any alternative.
The conclusion did not translate into a program. The Dow–Detroit Edison group’s own documents noted that the selection of the fast breeder was “made largely on the basis of the reactor’s ability to produce plutonium fuel as well as power and the value of plutonium to the government.” The weapons requirement was still bending civilian reactor development. Cisler’s engineers understood this, noted it clearly in their internal records, and went ahead with the fast breeder anyway because the government was willing to support it through the utility consortium structure that would become the Power Reactor Development Company.
A few weeks later, in April 1953, Cisler personally met with the AEC Commissioners to urge amendments to the Atomic Energy Act that would permit private industry to invest in nuclear projects. The Commissioners listened politely. They understood the logic. But the 1946 Atomic Energy Act gave the federal government a monopoly on fissile material, and in any case the AEC’s energies in those years were consumed by something else entirely.
The Arsenal Rises
Between 1950 and 1955, while the civilian power program inched forward and the first clearance-level surveys were being conducted by men like Cisler, the Atomic Energy Commission was executing the largest industrial expansion in American peacetime history. The scale of it dwarfed everything else the AEC was doing, and shaped — in ways that the civilian nuclear discourse of the 1950s almost never acknowledged — what American nuclear power would actually become.
The expansion was driven by three successive shocks. The first was the Soviet atomic test of August 29, 1949, which ended the American monopoly years earlier than the CIA had predicted. The second was Truman’s decision, on January 31, 1950, to authorize development of the thermonuclear weapon — the Super. The third was the North Korean invasion of South Korea in June 1950, which transformed the abstract confrontation with Soviet communism into a shooting war on the Asian continent.
But the shocks alone did not produce the response. Shocks of similar magnitude had produced, in other political contexts, cautious retrenchment rather than industrial expansion. What translated the shocks of 1949 and 1950 into the unprecedented buildout at Savannah River and Hanford was the specific political machinery of the Joint Committee on Atomic Energy under Brien McMahon, staffed by William Borden, and met at the AEC by a chairman — Gordon Dean — whom McMahon had personally installed. The triangle of those three men is the institutional channel through which the shocks flowed into concrete production programs.
Gordon Dean was forty-three years old in May 1949 when McMahon persuaded Truman to appoint him to the Atomic Energy Commission as one of two new commissioners. He was a lawyer, formerly of the Justice Department, most recently a professor at the University of Southern California Law School, with no background in atomic energy. His primary qualification was personal: he had been Brien McMahon’s law partner from April 1940 through the war, when Dean joined the Navy. McMahon and Dean were close friends, and McMahon had wanted Dean on the original AEC in 1946, only to be blocked by Lilienthal and Clark Clifford. In 1949, with Lilienthal wavering and a vacancy open, McMahon got what he had wanted for three years. Dean took his seat at the commission on May 24, 1949.
Fourteen months later, on July 11, 1950 — sixteen days after the North Korean invasion of the South — Truman named Gordon Dean chairman of the AEC. Lilienthal had resigned that February, exhausted by the fight over the Super. Truman had tried to draft Gordon Gray, had considered several other candidates, and had been opposed throughout by McMahon, who lobbied for Dean relentlessly. When the simultaneous expiration of all commissioners’ terms on June 30, 1950, forced Truman to make appointments at a moment when the Korean War had begun and Hickenlooper was threatening Pike’s renomination, McMahon brokered a package: Pike and Dean renominated together, Dean to the chairmanship. Truman agreed. On July 11, Dean had the job. Commission staff general manager Carroll Wilson, who believed Dean was McMahon’s tool, resigned three weeks later in protest and issued a public denunciation. McMahon collected eighteen signatures from the Joint Committee endorsing Dean within a half hour. The resignation made front-page news for a day and disappeared. Dean remained chairman.
McMahon’s staff chief was William Borden, the twenty-eight-year-old author of *There Will Be No Time*. Borden had been hired as Joint Committee executive director that same spring, 1949 — the year of the Soviet test, the year Dean joined the AEC, the year McMahon took back the Joint Committee chairmanship. From April 1949 through May 1953, the four most decisive years in the history of the American weapons complex, Borden occupied the most important staff position in congressional oversight of atomic energy. Dean’s office diary, kept by his secretary from notes she made while listening to his calls, records routine exchanges between the two men on everything from weapons tests to budget timing to public relations. Borden called Dean on matters as small as how to phrase a press release and as large as the structure of the second weapons laboratory. They were, during the Truman administration’s last three years, in daily contact.
What Borden and McMahon wanted, they had laid out in McMahon’s July 14, 1949 letter to Secretary of Defense Louis Johnson — a letter drafted by Borden. The letter argued that previous military requirements had “merely reflected an estimate of what the Atomic Energy Commission was capable of producing with existing or planned facilities — and did not reflect an independent judgment as to what we need in the event of war.” The existing program was too small, and too constrained by what the facilities could produce, rather than by what the military actually needed. McMahon and Borden believed the nation could never have enough atomic bombs. Strategic bombing with nuclear weapons was, in Borden’s phrase in the letter, “the keystone of our military policy and a foundation pillar of our foreign policy as well.” Everything followed from that premise.
The Soviet test of August 1949, six weeks later, gave the premise institutional force. Borden’s private assessment, prepared for the Joint Committee in the fall of 1949, was that “production goals were probably not based on the assumption of a Russian detonation in the summer of 1949.” No one had planned for the Soviet atomic monopoly to end in 1949. The AEC’s production schedule had been built around a longer American monopoly. Borden argued that the United States had, since the war’s end, devoted no more than one-thirtieth of its military budget to nuclear weapons. This was, in his view, absurdly inadequate. There should be, he wrote, “a substantial increase in the requirements for nuclear weapons and a new, concerted effort to develop the ultimate weapon system — the thermonuclear weapon carried by a nuclear-powered airplane.”
The argument led directly to the Super. The General Advisory Committee, meeting at Oppenheimer’s direction in late October 1949, had recommended against a crash program to develop thermonuclear weapons — arguing that the weapons were morally inappropriate, that their military utility was questionable, and that the effort would consume resources better spent on fission weapons. Lilienthal concurred. Conant, Rabi, Fermi, Oppenheimer — the most distinguished scientific body the American government had ever assembled — were opposed. Three out of five commissioners, led by Lilienthal, voted with the GAC.
McMahon’s reaction was not calm. He told Joseph Manley, the Joint Committee’s staff member who handled the weapons liaison, that he saw war with the Soviet Union as inevitable, that the GAC’s recommendation was “a suicidal response to a challenge by an immoral and implacable enemy,” and that the Super was “the only sure defense against such an enemy.” When Manley observed that this amounted to a “war-now” philosophy, McMahon accepted the characterization. The only alternative he could suggest was to announce as an ultimatum that the United States would proceed to develop the Super until the Russians “behaved.”
McMahon wrote a long letter to Truman on November 21, 1949, advocating the Super and refuting the GAC. His central claim was direct. “If we let Russia get the super first, catastrophe becomes all but certain — whereas, if we get it first, there exists a chance of saving ourselves.” Between November 1949 and January 1950, McMahon and Borden ran a public campaign — floor speeches, committee hearings, leaks to the Alsops, private appeals to Truman — that put the advisory committee’s recommendation under sustained political pressure. Teller and Lawrence supplied the technical case. McMahon supplied the political urgency. Strauss supplied the internal AEC vote.
Truman announced the decision to proceed with the Super on January 31, 1950. McMahon responded with a ringing Senate speech on February 2 assuring his audience that the President had made the right choice. The pattern of congressional pressure leading to presidential decision — Borden draft, McMahon speech, accumulated political weight, administration capitulation — was now established. It would be repeated, with variations, through the following two years.
The Soviet test had also generated pressure for more production capacity. Borden believed the existing reactor complex at Hanford was insufficient. On June 22, 1950, three days before the North Korean invasion, McMahon was reading a Borden-drafted classified report into the record of a Joint Committee hearing with General Electric officials. The report argued that the years 1952 and 1953 were “most often cited as years of greatest danger to the United States,” yet the President’s existing program would add nothing to the nation’s strength during this crucial period. Failure to build new reactors was, in Borden’s phrase, “subjecting the United States to grave peril.”
Then Korea. The North Korean invasion on June 25, 1950 — three days after Borden’s report — collapsed whatever remaining resistance there had been inside the administration. Truman approved two new heavy-water reactors on June 8; with the war, he authorized three more within months. The Savannah River Site emerged from a seventeen-site shortlist Du Pont had prepared; by late November 1950, Truman had signed a personal letter to Du Pont establishing Savannah River as a national defense priority, and construction was beginning. Du Pont, the same contractor that had built Hanford during the war, took a one-dollar fee and began work. By August 1951, Borden and Jackson were pressing for still more — “anything short of doubling the authorized output would be detrimental to the United States,” Jackson wrote to Secretary Johnson, with Borden’s analysis underneath the letter. McMahon was asking Dean and Johnson, on August 22, 1951, for their opinions on doubling the production rate by 1954.
By that fall the pressure was continuous and specific. McMahon in a December 1950 phone call told Dean: “I know you are fully conscious of living every day in a Pearl Harbor atmosphere, therefore, I would put myself, if I were you, in a pre-Pearl Harbor frame of mind. Where are you going to be if they hit tonight?” The diary records Dean’s reply — that the AEC had done exactly that on the stockpile — but the message was clear. From the Joint Committee’s chairman and staff director, Dean was receiving continuous reminders that he was commanding what amounted, in Borden’s framework, to the only force that could prevent national destruction.
The expansion programs that followed were sized accordingly. After preliminary skirmishing through 1951 over how large the next round should be, the National Security Council met on January 16, 1952, to consider the second major expansion. Lovett for Defense argued for it; Dean, more cautiously, supported the production increase while questioning its scale. Truman asked Lawton of the Bureau of the Budget a few questions, then paused. “In view of these considerations,” he said, “does anyone feel we should not undertake this?” There was no response. The program was approved. The new facilities authorized that day would expand uranium-235 production by 150 percent and plutonium production by 50 percent, enlarge Oak Ridge and Paducah, build an entirely new gaseous-diffusion complex at Portsmouth, Ohio, and add two new reactors at Hanford. The cost was measured in billions of dollars.
Dean himself, in subsequent private conversations, came to understand what was happening. His own internal memorandum in 1952 — drafted as the second expansion was being approved — observed that the new Savannah River reactors together with the new Hanford reactors would, within a few years, produce quantities of plutonium in excess of any conceivable military requirement. Tritium, with its 12.3-year half-life, would require permanent production capacity to replace natural decay. Plutonium-239, with its 25,000-year half-life, would accumulate indefinitely. Dean estimated the plutonium surplus would arrive “by the mid- or late 1960s.” Events would prove him right.
One passage in the Joint Committee’s November 1951 discussions with the Military Liaison Committee deserves particular notice. LeBaron was pressing Dean on how to maximize production. Someone around the table suggested that the Commission should obtain as much thorium ore as possible, given the alternative thorium-uranium fuel cycle that Weinberg and others at Oak Ridge had been exploring since the war. Dean’s reply was flat. “There were no plans to develop weapons using uranium 233.” The suggestion died where it was made. The thorium cycle was not useful for the weapons program, and the weapons program was what the Joint Committee cared about. Uranium-233 was not on the agenda, and would not get on it.
McMahon did not live to see the second expansion complete. He had been diagnosed with cancer earlier in 1952 and had been running for the Democratic presidential nomination — “Mr. Atom,” as the press had begun calling him, entering the field as a favorite son. From his sickbed in Georgetown Hospital that summer, he telephoned the Connecticut state Democratic convention to pledge that if elected president he would direct the AEC “to manufacture hydrogen bombs by the thousands.” He died on July 28, 1952, aged forty-eight. Truman paid tribute from Kansas City. Dean, in Vista, California, issued a statement calling McMahon a statesman of vision and energy, a good friend of the Commission, and a champion of world peace. Borden, at the Joint Committee offices, continued the work they had started.
The expansion that McMahon and Borden had built would outlive both of them in Washington. Borden’s tenure as Joint Committee executive director ended in May 1953, a few months after McMahon’s death, when Bourke Hickenlooper returned as Republican committee chairman and replaced him with Corbin Allardice. But the fourteen-reactor production complex — nine graphite reactors at Hanford, five heavy-water reactors at Savannah River — was under construction or operating. The gaseous-diffusion plants at Oak Ridge, Paducah, and Portsmouth were under construction or operating. The chemical separation plants, feed-materials facilities at Fernald, weapons-assembly facilities at Rocky Flats and Amarillo — the complete weapons production complex — was being built out on the timeline Borden and McMahon had demanded.
By the time Dean left the AEC on June 30, 1953, the American weapons production complex was essentially the complex that would operate through the 1980s. The construction costs ran into the billions of dollars. The operating workforce numbered in the tens of thousands. And the internal institutional knowledge — that the expansion would produce more plutonium than the weapons program could possibly use — had already been recorded, privately, by the man who had presided over its approval.
This is the single most important fact about the institutional context in which American civilian nuclear power took shape. By 1952 — before Shippingport was designed, before the Atoms for Peace speech, before the first commercial reactor order — the AEC’s own chairman understood internally that the plutonium economy was going to produce more material than the weapons program could absorb. The justification for fast-breeder development as a strategic necessity was already obsolete on its chairman’s desk. The case for the thorium cycle — which would not produce plutonium, and therefore would not feed the weapons complex — had been dismissed in a single sentence at the Military Liaison Committee table in November 1951.
But the justification persisted, politically, for another two decades. The institutional momentum of the expansion compounded faster than the strategic logic behind it could erode. Fourteen production reactors had been built, operated, maintained by Du Pont and General Electric and the Union Carbide contract-operators of the Oak Ridge plants. Thousands of engineers had been trained in the plutonium-separation chemistry the PUREX process required. Entire industrial ecosystems — the specialty alloys producers, the chemical-process contractors, the nuclear-fuel fabricators — had grown around the weapons program. When the time came to choose the direction of the civilian program, the weapons complex was the existing institutional reality, with its trained workforce, its established contractors, its familiar technology. The civilian program would be built on top of it, using its people, its contractors, its assumptions about what a nuclear reactor was.
Borden left the Joint Committee in May 1953, went to work for Westinghouse, and then, six months later, wrote a letter to J. Edgar Hoover accusing J. Robert Oppenheimer of being a Soviet agent. The letter was sent on November 7, 1953. The security hearing that followed — the public removal of Oppenheimer from government advisory service in April-May 1954 — would be the final act in the Borden program. The scientific establishment’s most prominent skeptic of continued weapons-complex expansion, the man who as GAC chairman had recommended against the Super in October 1949, was stripped of his clearance. The internal institutional opposition that had briefly delayed the Super, and might have delayed subsequent weapons-complex expansions, lost its most credible voice.
That was the meaning, in practice, of the institutional fusion that Eisenhower would later warn against.
The President’s Secret Briefing
Dwight Eisenhower had been elected president in November 1952. Before his inauguration, in December, a small meeting took place that almost no one knew about. Only Eisenhower himself and two AEC officials were present. The information conveyed was so sensitive that the AEC official who briefed the President-elect destroyed all his notes immediately after the meeting.
The subject was the nuclear weapons stockpile, which by late 1952 bore no resemblance to what Lilienthal had briefed Truman on five and a half years earlier. The JCS-authorized production schedule of April 1948 had committed the Commission to producing all the fissionable material its existing and under-construction facilities could manage. The post-1949 expansions had added the Savannah River complex, the new Hanford reactors, and the vast supporting chemical-industrial infrastructure. By the time Eisenhower took office, the weapons production program was consuming a substantial fraction of the total federal budget. The first thermonuclear device — the “Mike” shot, a wet, supercooled, building-sized device at Enewetak — had been tested only weeks earlier, on October 31, 1952, with a yield of ten megatons. Its successors would be smaller, deliverable, and, with the discovery of the Teller-Ulam design in 1951, capable of arbitrary yield.
The figures Eisenhower received that December were, for him, disturbing. He had not been a pacifist; as supreme commander of Allied forces in Europe, he had overseen the most destructive military operation in history. But he had also, more than most American politicians, watched close at hand what war did to societies. He had seen Germany’s cities. He had seen the refugees. And he understood, because he had been briefed on the physical effects of thermonuclear weapons, what a full-scale nuclear exchange would mean. The numbers presented to him in December 1952 — a stockpile measured in thousands, growing into tens of thousands, with thermonuclear weapons whose yields dwarfed Hiroshima — troubled him in a way that the public Eisenhower, the confident soldier-statesman, never revealed.
His response, over the following months, was complicated and somewhat contradictory. He did not try to halt weapons production. He did not cut the AEC’s budget for fissile material. He accepted, and in some areas accelerated, the weapons program that Truman had authorized. But he also looked for a way to counter the psychological weight of the nuclear arms race with a public gesture toward peaceful use.
That gesture became Atoms for Peace.
Atoms for Peace
On December 8, 1953, Eisenhower stood before the United Nations General Assembly in New York and delivered one of the most remarkable speeches of the early Cold War. He acknowledged the destructiveness of nuclear weapons in plain terms. He described the American stockpile as adequate to the nation’s defense but as also, in its growth, a danger to humanity. He proposed the creation of an International Atomic Energy Agency, to which the nuclear powers would contribute fissile material for peaceful purposes — reactors, research, medical applications — and which would distribute that material to nations seeking to develop civilian nuclear programs.
The proposal was vague enough that its details could be argued about for years, and they were. But the speech itself was electric. Lewis Strauss, who had succeeded Gordon Dean as AEC chairman the previous summer, later described the scene. The delegates sat in silence through the twenty-minute address. When Eisenhower finished, even the Soviet delegation joined the applause. For a brief moment in the middle of the Cold War, the promise of civilian nuclear power seemed to offer a path out of the weapons race.
What Atoms for Peace actually committed the United States to was something more specific, and more consequential for the civilian reactor program, than most listeners understood. Eisenhower had promised to share nuclear technology with allied nations. That required fissile material for research and power reactors abroad. It required revised legislation that would permit American industry to own such reactors and sell them to foreign customers. And it required, most critically, a domestic civilian nuclear industry robust enough to supply the international demand.
In other words, Atoms for Peace required a civilian power reactor program that did not really exist yet. It required, in fact, that the United States demonstrate civilian nuclear power on American soil, as quickly as possible, so that the technology could be exported with confidence.
The Commission had already been thinking about this. Earlier in 1953, in response to a budget crisis and the new administration’s desire for economies, the Navy’s nuclear aircraft carrier reactor had been cancelled. Rickover and Thomas Murray, an AEC commissioner, had proposed that the work Westinghouse had already done on the pressurized-water reactor for the cancelled carrier be redirected toward a civilian power plant. The proposal found a receptive audience. On May 6, 1953, the National Security Council approved using the carrier reactor work for a new central-station nuclear power plant, at a total cost of $100 million, “unless private financing should become available before completion.”
Private financing did not become available. Private industry, as Cisler and others had made clear, was not in a position to absorb the risk of a first-of-a-kind civilian power reactor. The project would be a government undertaking, contracted to Westinghouse through the AEC’s Naval Reactors Branch, under Rickover’s personal direction.
The decision to put Rickover in charge of the civilian project was not obvious. Strauss had initially questioned whether a reactor led by Rickover would be credibly “civilian.” Some members of the AEC’s reactor development staff argued that the proposed plant was neither large enough nor novel enough in design to demonstrate nuclear power’s commercial potential. Some electric utility executives wanted industry to have more control. Murray resolutely countered these arguments. The AEC approved Rickover’s leadership in July 1953 and announced it publicly in October. Westinghouse and the Bettis laboratory were already at work.
The civilian project was named for its location: a site on the Ohio River near Pittsburgh, in the small town of Shippingport.
The Aircraft Reactor
On the afternoon of November 3, 1954 — three and a half years before Shippingport achieved criticality — a small, loud, irregular clicking began in the control room of a nondescript concrete-block building at Oak Ridge National Laboratory. The building was in a part of the lab’s site that had not been used during the war. The clicking came from three neutron detectors: two fission chambers in the reflector of a reactor called ARE, and a BeF₃ counter positioned outside the core. The count rates had been rising all week as the operators added successive aliquots of uranium fluoride concentrate to the fuel circuit. At 3:45 PM on November 3, upon the completion of the twelfth enrichment operation, the rates stabilized and began to rise of their own accord. The reactor had gone critical. Molten salt at nine hundred and fifty degrees Fahrenheit was circulating through the core, fission was taking place in the liquid fuel, and the thermal output was — at the moment of criticality — essentially zero, the reactor operating at watts rather than kilowatts or megawatts. But the physics was working. The first molten-salt reactor in history was alive.
The Aircraft Reactor Experiment had a strange institutional origin, and that origin is worth understanding because it shapes everything about what the ARE was and was not allowed to become. The reactor was not built to generate civilian power. It was built to produce data for a program — the Aircraft Nuclear Propulsion program — whose premise was that the United States needed long-range bombers capable of transoceanic flight without refueling, that conventional fuel limited bomber range, and that only a reactor small enough to fly could carry the energy density required. Curtis LeMay had championed the concept inside the Air Force beginning in 1946. By the late 1940s the Air Force had convinced itself that a nuclear-powered bomber was strategically essential, and the Atomic Energy Commission, after some institutional maneuvering, had taken responsibility for the reactor development portion of the program. The original contractor, Fairchild, had started work at Oak Ridge in 1946 under the acronym NEPA — Nuclear Energy for the Propulsion of Aircraft — and had failed to make meaningful technical progress for several years. In 1951, the AEC reorganized the program, bringing the reactor work under Oak Ridge National Laboratory’s direct management. R. C. Briant was named project director. It was at that point, and under that institutional cover, that the molten-salt reactor program began in earnest.
Briant was the critical figure in what followed. The Aircraft Nuclear Propulsion program required a reactor that could operate at temperatures high enough to produce useful thrust from a jet engine — 1,500 degrees Fahrenheit at minimum, preferably higher — and at pressures low enough to be carried aloft in an aircraft without catastrophic failure under combat conditions. No solid-fueled reactor could meet both requirements simultaneously. At 1,500 degrees, solid fuel pins would fail from xenon instability and thermal stress; the cladding metallurgy did not exist. Pressurized water reactors of the sort Rickover was building for the Navy operated at 550 degrees and would not produce jet-engine thrust. The aircraft program needed a different class of reactor entirely.
Briant’s insight, developed in consultation with Alvin Weinberg and with Eugene Wigner’s earlier thinking in mind, was that a fluid fuel could meet both requirements at once. Molten fluoride salts — specifically, mixtures of uranium tetrafluoride with the fluorides of sodium, zirconium, beryllium, or lithium — had melting points below 1,000 degrees Fahrenheit and boiling points well above 2,000. They were radiation-stable, dissolved uranium at the required concentrations, and could be pumped through a reactor core at temperatures a solid-fuel system could not approach. If the chemistry worked and the materials could be found to contain the salt, a molten-salt reactor could operate at 1,500 degrees or higher and at nearly atmospheric pressure. The aircraft program had a technical requirement that only a fluid-fuel reactor could satisfy. The molten-salt concept, which Wigner and Weinberg had proposed in 1945 for civilian power generation, found its first actual funding through the back door of a bomber program.
E. S. Bettis led the design. The work began with a solid-fuel reactor — uranium oxide pins in stainless steel cladding, moderated by beryllium oxide blocks, cooled by circulating sodium. The beryllium oxide blocks had already been procured before the project changed direction. When xenon instability and thermal-stress calculations killed the solid-fuel approach in the first design phase, the beryllium oxide remained, and the team pivoted to a stagnant liquid-fuel design that used the same moderator. The stagnant design failed in turn — the thermal gradient across a tube of stagnant molten fluoride, at practical tube diameters, pushed the centerline fuel temperature dangerously near the salt’s boiling point. Only a circulating fuel, pumped through the moderator at turbulent flow rates, would carry the heat out fast enough to keep the core safe. The final design, assembled from this sequence of failures, was what reached criticality in November 1954: a cylindrical Inconel pressure vessel containing hexagonal beryllium oxide moderator blocks, threaded with sixty-six Inconel fuel tubes arranged in six parallel circuits of eleven passes each, through which a NaF-ZrF₄-UF₄ fuel salt was pumped at fourteen thousand gallons per hour.
The reactor was installed in three concrete pits below the basement floor of the ARE building, with 7.5 feet of concrete blocks serving as biological shielding when the pits were closed. Four sleeves penetrated the core for control: three shim rods and a central regulating rod, all stainless steel driven from above by electric motors. The fuel system was non-drainable — a design defect forced by the beryllium oxide moderator geometry — which meant that once the fuel was loaded, it had to stay in place until the experiment concluded. Every step of the operation had to work the first time, because nothing could be undone. The team understood the hazard and proceeded with what the official operating history would later describe as “rather drastic safety features” and “extreme care in every phase.”
The fluoride carrier salt, NaF-ZrF₄, was melted and transferred into the fuel system on October 28, 1954. Uranium fluoride concentrate — enriched to 93.4 percent ²³⁵U — was added in small weighed quantities from portable cans through an intermediate transfer pot. Each addition was calculated to bring the reactor closer to criticality; each addition’s effect on the count rates was measured; and the projection curve was updated. Three days of enrichment operations followed, interrupted by repeated blockages of the concentrate transfer line and occasional leaks that had to be cleared. At the twelfth enrichment, on November 3, the reactor went critical with roughly sixty-one kilograms of ²³⁵U in the system, though only fourteen-point-nine kilograms of that was actually in the core at the operating temperature. The rest was in the piping, the pump bowl, and the heat exchangers — fuel that was part of the circuit but not part of the critical mass.
Over the following five days, the team performed low-power experiments to calibrate the instrumentation and measure the temperature coefficient of reactivity. The temperature coefficient was the first critical physics question: would the reactor respond to rising temperature by generating less power, or more? If it responded with more power, the reactor would be unstable, and shutting it down in an emergency would be difficult. Weinberg had predicted that the molten-salt reactor would have a strongly negative temperature coefficient, because the salt expanded substantially as it warmed, and expansion pushed fuel out of the core region — reducing the reactivity directly. The low-power measurements confirmed the prediction. The reactor was inherently stable.
On the evening of November 9, 1954, the reactor went to one megawatt of thermal power. Over the next three days it ran intermittently at levels up to 2.5 megawatts thermal, with a fuel outlet temperature at steady state of 1,580 degrees Fahrenheit — achieving, experimentally, the temperature regime the Aircraft Nuclear Propulsion program required. The measured reactor temperature coefficient of reactivity was –6 × 10⁻⁵ Δk/k per degree Fahrenheit, consistent with theory and strong enough to make the reactor, in Weinberg’s later phrase, “a slave to the power load.” If the heat extraction was increased — by speeding the helium blower that cooled the fuel heat exchanger — the reactor’s fuel cooled, contracted, pulled more neutrons into the core region, and the reactor power rose to match the new extraction rate, with no rod movement required. If the extraction decreased, the reactor automatically throttled back. The control rods determined the mean operating temperature, but not the power level. The power level was set by the load.
One finding from the ARE’s operation was consequential beyond anything the Aircraft Nuclear Propulsion program needed. After twenty-five hours at full power, the team measured the xenon-135 buildup in the core — the fission product whose neutron-absorption cross-section is a hundred times that of any other isotope, and which in solid-fueled reactors forces costly shutdowns every few weeks because it poisons the chain reaction. In the ARE, no xenon-135 poisoning could be detected. As little as five percent of the theoretically expected poisoning would have been measurable; none was found. The xenon had simply bubbled out of the circulating fuel at the pump bowl, where the liquid interface with the gas space allowed the volatile fission product to escape continuously. The circulating-fuel architecture was self-cleaning of its worst poison. The reactor could run indefinitely at full power without the refueling shutdowns that defined the operating economics of every solid-fuel reactor then under design.
The Aircraft Reactor Experiment was shut down on November 12, 1954, after a total of 221 hours of nuclear operation and 96 megawatt-hours of energy produced. Seventy-four of those hours had been in the megawatt range. The fuel and sodium circuits had operated for 462 and 635 hours respectively, at temperatures no solid-fuel reactor could approach. The experiment had been completed on schedule. Every nuclear measurement planned had been achieved. Weinberg was, years later, explicit about what the ARE had proven: that a fluid-fueled reactor could be built, that it would be inherently stable, that its fuel chemistry could be managed, and that the class of materials problems the circulating molten salt presented could be solved. “The ARE,” he wrote, “showed that the concept was sound.”
What it did not do was save the Aircraft Nuclear Propulsion program. The strategic case for a nuclear-powered bomber was already weakening by 1954. The successful hydrogen bomb tests of 1952 (Mike, ten megatons) and 1954 (Castle Bravo, fifteen megatons) had demonstrated that an intercontinental delivery system carrying even one or two warheads could destroy an adversary regardless of the delivery aircraft’s fuel capacity. The Strategic Air Command’s B-52, entering service in 1955, had sufficient range with air-to-air refueling to reach any target the United States cared about. The first ICBM tests followed shortly thereafter, and by the end of the 1950s the intercontinental ballistic missile had rendered the nuclear-powered bomber strategically redundant before any airframe had flown. Eisenhower questioned the program throughout his administration. Kennedy cancelled it in March 1961. The Air Force’s nuclear-powered bomber never existed as anything but blueprints and mockups.
But the ARE existed. It had run. The reactor that had demonstrated the molten-salt concept was a working piece of nuclear engineering, and its lessons did not die with the program that had funded it. Weinberg and Briant had used the aircraft program’s money and institutional shelter to prove, at multi-megawatt scale, the physics and chemistry of a civilian power reactor technology that the AEC’s civilian reactor development program had not, on its own authority, been willing to fund at that level. The Oak Ridge team had built a reactor ostensibly for a bomber and had, in effect, built the first demonstration version of a thorium breeder. The distinction between the two — what the reactor was called and what it actually demonstrated — would structure the next decade and a half of Oak Ridge’s fluid-fuel work.
The Molten Salt Reactor Experiment that would be built at Oak Ridge beginning in 1962, criticality in 1965, uranium-233 operation by 1968 — that reactor would descend directly from the ARE. Its engineers were, in many cases, the same engineers. Bettis, MacPherson, Briant before his death in 1961, and a team of Oak Ridge metallurgists, chemists, and reactor physicists that grew through the late 1950s into what was, by the mid-1960s, the world’s concentration of molten-salt reactor expertise. When the AEC’s Fluid Fuel Reactors Task Force would meet in January-February 1959 and rank the molten-salt concept as the most promising of the three fluid-fuel technologies, the single most important operational evidence before them was the ARE’s 1954 run. When the Geneva 1958 conference disclosed the molten-salt concept to the world in *Fluid Fuel Reactors*, H. G. MacPherson’s Part II included a dedicated technical chapter on the ARE’s operation. The civilian molten-salt program that would emerge from the late 1950s was not a new program. It was the continuation of what the ARE had started.
Back in the building at Oak Ridge where the ARE had operated, the concrete shielding blocks were left in place over the pits for years. The reactor vessel, still containing frozen fluoride fuel that had been drained at the end of the experiment and allowed to solidify, was never disturbed. The ARE building became, over time, a quiet corner of the laboratory — one more structure among many on a sprawling site. But the data from its 221 hours of nuclear operation, written up in three papers for *Nuclear Science and Engineering* in 1957 by Bettis and his colleagues, would remain the foundational experimental record for every molten-salt reactor program attempted anywhere in the world for the next seventy years.
The Submarine Reactor
In the sparse desert eighty miles west of Idaho Falls, on the floor of a great bowl ringed by the jagged peaks of the Lost River Range, stood in the spring of 1953 a strange installation. A large rectangular sheet-metal building — unadorned, utilitarian, obviously industrial — contained what looked like a section of an enormous tunnel that had been dug up from somewhere else and placed indoors. Inside that tunnel, against curved vertical plates that held a water tank, was a mockup of a submarine’s pressure hull. Inside the hull was a reactor. Around the reactor was the steam plant a nuclear-powered submarine would carry to sea. All of it was on dry land, a thousand miles from the nearest ocean, surrounded by sagebrush and dust. It was the Mark I submarine thermal reactor, and it was the first fully engineered nuclear reactor capable of producing practical amounts of power on a sustained and reliable basis. It was, in that narrow engineering sense, the piece of machinery that was about to change the course of the civilian nuclear age.
The reactor had been conceived by a man who was, in the spring of 1953, about to be forced to retire from the United States Navy as a captain who had twice been passed over for promotion to rear admiral. Hyman Rickover had been told he was too abrasive, too specialized, too narrow in his interests to be given flag rank. On March 30, 1953, with the Mark I reactor approaching first criticality in the Idaho desert, Rickover was in the control room. The operators had been inching the control rods out of the core for hours, checking the hundreds of instruments monitoring every part of the plant, stopping repeatedly when a safety circuit tripped the automatic shutdown and made the team begin the analysis over again. Just before midnight — at 11:17 PM local time — the neutron counts stabilized and began to climb. Mark I had gone critical. Its power level at that moment was less than a hundredth of a horsepower, sufficient for physics data and shielding measurements and nothing else. But the reactor was alive. The second reactor in the United States to achieve criticality as a practical engineering device — after the Aircraft Reactor Experiment in November 1954, which Rickover’s reactor had preceded by sixteen months — was working.
The path Rickover had taken to this result had been unusual, though not in retrospect unreasonable. When Rickover had gone to Oak Ridge in 1946 as one of the Navy’s six officers sent to learn what nuclear reactors were, he had concluded fairly quickly that the fluid-fuel concepts Weinberg and Wigner were pursuing were not, yet, reactors. They were research programs. For a submarine — a vessel that had to run submerged for months at a time, in spaces too small for extensive auxiliary equipment, with operators who could not be expert reactor physicists — calculable was essential. A submarine reactor had to be reliable the way a ship’s boiler was reliable. Rickover had concluded that pressurized water, using ceramic uranium dioxide fuel pellets sealed in zirconium cladding, was the only technology he could specify to his engineers and expect to get a working reactor in five years. It was not the best reactor for civilian power. It was the reactor he could build on the schedule he had.
Having made that decision, Rickover had made another that was less obvious but equally consequential. He had rejected what every other reactor development program was doing, which was to build a test-stand “breadboard” reactor first, then a shipboard prototype, then the operational plant — each step based on the successful completion of the previous one. Rickover had instead pushed for concurrent development. The Mark I would be a land-based prototype whose reactor and steam plant were arranged exactly as they would be on an actual combat submarine. The Mark II, which would go into the actual submarine, would be built in parallel with the Mark I — not after it. The submarine’s keel would be laid before the prototype had even begun to operate. Rickover called this strategy “Mark I equals Mark II.” Everyone at the Bureau of Ships who heard it understood that it was a gamble. If the Mark I encountered a fundamental obstacle — anything in the laws of nature that made the whole approach unworkable — the submarine, already under construction, would be an expensive piece of scrap metal. And the responsibility would be Rickover’s personally.
Construction had begun in Idaho in the spring of 1950, with Commander Jack McGaraghan from the Navy’s civil engineer corps supervising work on the site. Westinghouse at Bettis Laboratory near Pittsburgh was designing the reactor and steam plant components; Electric Boat in Groton, Connecticut, was building the hull section that would house them. The level of administrative discipline Rickover imposed on the work was, even by the Navy’s standards, unusual. Every component had to be fabricated under cleanliness standards normally reserved for pharmaceutical manufacture. Every piece of piping had to be tagged, wrapped, inspected, logged. McGaraghan had to write Rickover a “gripe” letter every week that went to Rickover personally. Weekly summaries of production meetings at Bettis were teletyped to Idaho. John O’Grady, a production-control officer Rickover had hired from the wartime Navy, spent his time badgering suppliers on delivery dates and once physically extracted reactor components from a strike-bound Babcock & Wilcox plant.
Alongside the reactor work, Electric Boat was building a full-scale mockup in wood and cardboard of every pipe, valve, electrical panel, and major piece of equipment that would go into the actual submarine’s reactor and machinery compartments. Rickover would climb through these simulated compartments on his visits to Groton, looking at the placement of valves and handwheels from every angle, making sure that a maintenance task could be performed at sea by a crewman in uniform without unusual difficulty. The Mark I itself was being built in Idaho as though it were a ship — because in the concurrent-development philosophy, the engineers who built the Mark I would be the same engineers who would later fit out the Nautilus at Groton. The distinction between prototype and operational plant was being deliberately collapsed.
The reactor went critical on schedule in March 1953. Over the following two months, as the operators mapped the physics of the cold clean core, Rickover flew Thomas Murray — the AEC commissioner who was an engineer by training and had become Rickover’s most reliable supporter on the Commission — to Idaho to witness the first steam. On May 31, 1953, Murray opened the throttle. Nuclear-generated steam entered a turbine for the first time in history. The Mark I produced several thousand kilowatts of shaft power. It worked.
What followed was the most consequential test in the history of American naval technology. Rickover had planned, as a demonstration of sustained operation, a forty-eight-hour full-power run. Then his physicists had come back to him and said that twenty-four hours would yield all the data they needed. Rickover, on a plane to Idaho, learned of the change and countermanded it. He made it a one-hundred-hour run. He told his people that he wanted to test not only the nuclear components but the steam plant, the turbines, the coolant pumps — everything that would be at sea. Then, on a cot in a Quonset hut at the site that night, Rickover thought of something else. He told the crew in the control room to post on the wall a chart of the North Atlantic and, at the end of every four-hour watch, to mark the position of the “ship.” The Mark I was not a ship. It was a reactor in a sheet-metal building a thousand miles from the nearest ocean. But its steam was driving a turbine connected to a dynamometer, its crew was on four-hour submarine watches, and the chart of the North Atlantic would be the measure of what the reactor was doing.
The run began in late June 1953 and continued for four days. Around the sixtieth hour, motor-generators in the steam plant began to act erratically. A reactor coolant pump developed strange noises. The nuclear instrumentation became unreliable. From Washington, where he had returned partway through the run, Rickover refused every request to shut down. The problems were in the steam plant, not the reactor; the reactor was operating within its design envelope; nothing was wrong that could not be monitored and managed. In the control room at the site, the officers kept marking positions on the chart every four hours. When their reckonings showed the “ship” had reached the coast of Ireland, they shut down Mark I in accordance with planned procedures. It had run for one hundred hours at full power. It had, in engineering terms, crossed the Atlantic.
The 100-hour run was the moment the civilian nuclear age effectively began. The Mark I became, that summer, the world’s first fully engineered nuclear reactor capable of delivering practical amounts of energy on a sustained and reliable basis. Everything that had happened before — the Aircraft Reactor Experiment’s 2.5-megawatt run in November 1954, still sixteen months in the future; the Homogeneous Reactor Experiment’s early tests at Oak Ridge; the various experimental piles at Argonne — had been research. Mark I was engineering. It was the proof that a specific reactor design, with specific cladding materials and coolant chemistry and component designs, could run reliably for the sustained periods that any commercial or naval application required. The utilities that would eventually buy civilian reactors would buy them on essentially the basis the Mark I had established.
Rickover’s personal situation changed the same summer. In March 1953, just days before the Mark I achieved criticality, Senator Henry Jackson had intervened in the Navy’s promotion proceedings by asking Senator Saltonstall’s Armed Services Committee to withhold action on thirty-nine captain-to-admiral promotions pending an investigation of the selection system that had twice passed over Rickover. Congressman Sidney Yates had delivered the preparatory salvos from the House floor in February. The Navy, facing the prospect of having all its rear admiral promotions for 1953 delayed by a congressional fight over one engineer, capitulated. Secretary Anderson convened a special board. In July, after the Mark I’s successful 100-hour run, Rickover was promoted to rear admiral. He would remain head of the Naval Reactors Branch — the only position he ever wanted — for the next twenty-nine years, the longest-serving active-duty officer in American history by the time of his eventual retirement in 1982.
While the Mark I was running in Idaho, the Mark II — the actual shipboard reactor — was taking shape at the Electric Boat yard in Groton. The keel of the Nautilus had been laid on June 14, 1952, by President Truman personally; Rickover stood behind the president among a cluster of dignitaries for the ceremony. By the time the Mark I achieved criticality, the submarine’s hull was being assembled around a reactor compartment that was, as far as physical arrangement was concerned, identical to what Rickover had built in the Idaho desert. The concurrent-development gamble had worked. The submarine was being built to the same drawings as the prototype.
The Nautilus was launched on January 21, 1954 — hull complete, reactor compartment ready for installation of the Mark II. Through 1954, Electric Boat and Westinghouse engineers worked around the clock to finish the reactor installation and complete the ship. In September a serious problem surfaced: faulty steam piping that had to be ripped out and replaced. The schedule Rickover had fought for — a January 1, 1955 delivery — was in danger. The team compressed every step of the remaining work. On December 30, 1954, the Mark II reactor achieved criticality at the pier in Groton. On January 2, 1955, the reactor briefly supplied steam to the ship’s electrical system. Later that day, steam was fed to the turbines, and the propellers turned for the first time while the ship lay at the pier. On January 3, the Mark II reached full power.
Two weeks later, on the morning of January 17, 1955, the Nautilus dropped its mooring lines at the Electric Boat pier. Wilkinson — Commander Eugene Wilkinson, the ship’s captain, whom Rickover had personally chosen over other candidates — gave the command to back. Rickover was on the bridge beside him. As the ship pulled away from the dock, the engineering officer in the maneuvering room reported a loud noise in the starboard reduction gear and switched to electric propulsion. Under normal circumstances a captain would have returned to the pier. But press boats were in the channel and Rickover was not inclined to abort the trial. Paul Panoff and the engineering officer went below, found that a locking pin on a retaining nut had come loose, and replaced it in a few minutes. The ship shifted back to steam propulsion. The Nautilus passed down the Thames River and out past the breakwater into Long Island Sound. A signalman blinked a message to the escort tug Skylark: UNDERWAY ON NUCLEAR POWER.
The sea trials that followed exceeded what the Navy had hoped for. The shakedown cruise took the Nautilus from New London to San Juan, Puerto Rico — 1,300 miles submerged, in eighty-four hours, at sustained speeds around sixteen knots. No submarine had ever traveled one-tenth that distance submerged before, or maintained that speed for more than an hour. The Nautilus had broken, in a single shakedown cruise, the fundamental operational constraint that had defined submarine warfare since the Holland boats of 1900. A diesel-electric submarine submerged could run at full battery discharge for perhaps an hour, or at low speed for perhaps twenty-four; after that it had to surface to run diesels and recharge batteries. The Nautilus simply did not have to surface. Its nuclear reactor produced steam as long as the reactor ran, and the reactor would run for years between refuelings.
The consequences became visible quickly. In July and August 1955, the Navy conducted an anti-submarine exercise in which a task force of destroyers and a carrier with its attack aircraft tried to locate and attack the Nautilus operating alongside several conventional Guppy-class diesel-electric submarines. The results were one-sided. The Guppies were hard pressed. The Nautilus was, in the assessment of the exercise evaluators, essentially invulnerable. It could locate the hunter-killer group at long ranges from its sonar. The surface force could not detect the Nautilus. It could outrun any torpedo the defenders had. It could close, attack, and disengage without ever revealing its position. Admiral Jerauld Wright, Commander in Chief of the Atlantic Fleet, wrote after reading the final report: “It is urgent that countermeasures be developed for the true submarine and that no future combatant submarine be built that is not nuclear powered.”
Three things followed from the Mark I-Nautilus program that would define the American civilian nuclear industry for the next fifty years, though almost no one in the nuclear industry understood them clearly at the time.
The first was a specific reactor architecture. Pressurized water at roughly 2,000 pounds per square inch, ceramic uranium dioxide fuel pellets at low enrichment sealed in zirconium cladding, a secondary steam loop to drive the turbine, control by neutron-absorbing rods and by boric acid dissolved in the primary coolant — this combination of design choices had been made, not because it was optimal for civilian power generation, but because it could be built for a submarine on the schedule and in the space Rickover had. The light-water reactors that American utilities would order through the 1960s and 1970s — Shippingport first, then Yankee Atomic, Dresden, Indian Point, and eventually the hundred-plus commercial reactors that would be built in the United States — all descended from this naval-reactor template. The ceramic fuel pellets were the same. The zirconium cladding was the same. The pressurized primary loop was the same. The control approach was the same. The engineering knowledge base that allowed a utility to order a reactor from Westinghouse in 1965 was the knowledge base Bettis Laboratory had built for the Mark I in 1950-1953.
The second was a specific industrial infrastructure. Westinghouse, the contractor Rickover had chosen for the submarine reactor, built its commercial PWR business on the engineering staff, the metallurgical expertise, and the manufacturing capability it had developed at Bettis. General Electric, which had lost the Navy’s sodium-cooled intermediate reactor competition and had watched its Seawolf Mark A plant be removed from service after unreliable operation, pivoted to its own commercial boiling-water reactor — which was similar enough to the pressurized-water design in its core architecture to share the same fuel and cladding industrial base. The nuclear-equipment industry that would spring up to support the utility orders of the 1960s — the fuel fabricators, the pressure-vessel manufacturers, the steam-generator suppliers, the zirconium-alloy producers — was, to a remarkable degree, an extension of the industrial network Naval Reactors had built during the Mark I-Nautilus period. By the time the first commercial reactor orders arrived, the supply chain already existed. It was Rickover’s supply chain.
The third was a specific engineering profession. Rickover had personally trained the officers and crew of the Nautilus through a year of mathematics, physics, and reactor engineering at Bettis before they ever reported aboard. Every subsequent commanding officer of a nuclear-powered Navy ship passed through the same training pipeline, personally screened by Rickover through the end of his tenure. The civilian nuclear industry, as it grew through the 1960s, drew its senior engineering talent — at first haltingly, then massively — from the officers Rickover had trained and mustered out of the Navy. The engineering culture of American civilian nuclear power was shaped, in its assumptions about quality control, documentation, training, and personal accountability, by the culture Rickover had imposed on his submarine program. This was not an accident. It was Rickover’s deliberate choice, sustained over decades, to make the civilian industry safe enough that it would not discredit the nuclear Navy.
The Mark I prototype in Idaho, which had crossed an imaginary Atlantic in the summer of 1953, was eventually shut down in 1989 after thirty-six years of use as a training reactor for Navy officers. Its reactor vessel was removed and buried. The sheet-metal building in the desert was demolished. The Nautilus herself was decommissioned in 1980 and moored at the Submarine Force Museum in Groton, Connecticut, where she remains as of this writing — a 320-foot pressure hull with a brass plaque on her bridge that records the message blinked to the Skylark on the morning of January 17, 1955. But the architecture the Mark I had established — the pressurized-water reactor, the fuel form, the industrial suppliers, the engineering training pipeline, the entire culture — was still running, fifty years later, at scores of civilian plants whose designers had, in most cases, never been inside a submarine.
The alternative technologies that the Fluid Fuel Reactors Task Force would rank, in February 1959, as more promising than the light-water reactor for civilian power generation — the molten-salt reactor that Weinberg and MacPherson were developing at Oak Ridge, the liquid-metal-fueled reactor that Brookhaven and Babcock & Wilcox were developing — would never be built at comparable scale. They would never have their own Mark I. They would never cross their own imaginary Atlantic. And by the time the civilian industry had matured to the point of asking whether it had chosen the right reactor, the answer was no longer one the industry could easily change. The reactor it had was the reactor Rickover had built for a submarine.
Shippingport
The Shippingport Atomic Power Station was, in almost every respect, a scaled-up naval reactor.
The Westinghouse team at Bettis had spent the preceding five years designing and testing submarine reactors under Rickover’s direction. They had built the S1W land-based prototype in the Idaho desert. They had worked out the metallurgy of zirconium cladding, the chemistry of high-temperature water as both coolant and moderator, the geometry of ceramic uranium-dioxide fuel pellets sealed in Zircaloy tubes. All of this knowledge was transferable, directly, to a stationary power reactor. What Shippingport would demonstrate, in essence, was that a submarine-derived reactor could be made larger and configured to produce electricity for a utility grid.
It would not demonstrate that this was the best way to generate civilian nuclear power. It would demonstrate that it was a way — the way that Rickover’s team knew how to build, within the time the political situation demanded.
Construction began in 1954. Ground was broken in September with a Rickover-style ceremony: brief, unsentimental, heavy on technical detail. Eisenhower himself triggered the initial construction remotely, waving a neutron wand over a radiation detector at a conference in Denver. The reactor achieved criticality on December 2, 1957, exactly fifteen years after Fermi’s original pile under the stands of Stagg Field in Chicago, and began generating power into the Duquesne Light Company grid on December 18.
It worked. It worked reliably. Over the following twenty-five years, Shippingport would operate at various power levels and with several different core configurations, including, eventually, a light-water breeder core using the thorium-U233 cycle that Rickover personally championed in the 1970s. As a technical demonstration, it was an unambiguous success.
As a template for the American civilian nuclear industry, it was something else. The utilities that followed Shippingport into nuclear power — Yankee Atomic, Consolidated Edison, Northern States Power — built variants of the pressurized-water reactor. The General Electric boiling-water reactor, an alternative that eliminated the secondary steam loop at the cost of allowing primary coolant into the turbine, was the only major deviation from the basic Westinghouse architecture. Both types were direct descendants of Rickover’s submarine work. Both used solid fuel elements. Both operated with water as primary coolant. Both accumulated xenon-135 in their cores and required the same elaborate refueling shutdowns that B-Reactor had taught the Hanford engineers to expect in 1944.
By late 1957, Shippingport was operating, the *Nautilus* had sailed under the polar ice cap, Sputnik had launched, and American utilities were beginning to discuss in earnest the prospect of commercial nuclear orders. To anyone reading the general press in the United States, the template of the civilian nuclear age looked settled: water cooled, solid fueled, Rickover-derived.
And to anyone reading the general press, that was the whole story. It was not the whole story.
Geneva 1958
In September 1958, three thousand delegates from sixty-eight nations gathered at the Palais des Nations in Geneva for the Second International Conference on the Peaceful Uses of Atomic Energy. It was the largest technical conference the postwar world had yet seen. The first Geneva Conference, three years earlier, had been the occasion for the initial wave of American nuclear-technology declassification — the moment when reactor physics, which had been classified since the Manhattan Project, became openly discussable. The Second Geneva Conference was the moment of fuller disclosure. Detailed engineering designs, materials data, cost studies, operating experience from experimental reactors — all of it was opened to the international community. The AEC, under Strauss’s chairmanship, treated Geneva 1958 as Atoms for Peace made manifest. The United States had promised to share. This was the sharing.
Among the thousands of papers presented at Geneva were contributions from a book-length technical compendium prepared that summer under AEC auspices and intended for open publication. Its title was *Fluid Fuel Reactors*. It ran to nine hundred pages. It was organized in three parts, each edited by a leader of one of the three major American fluid-fuel reactor programs. It described, openly, to the world, what the American national laboratories had been doing in parallel with the Shippingport program.
Part I, edited by James A. Lane of Oak Ridge National Laboratory, covered aqueous homogeneous reactors — the fluid-fuel concept originally developed by Weinberg’s group in the late 1940s, in which uranium was dissolved as uranyl sulfate in heavy water, circulated continuously, and processed to remove fission products. The Homogeneous Reactor Experiment, HRE-1, had operated at Oak Ridge from 1952 to 1954. Its successor, HRE-2, had reached criticality in December 1957 and was operating at five megawatts. The aqueous homogeneous program represented, by 1958, more than ten years of research and development, some 1,500 technical man-years of effort, and more than fifty million dollars of federal investment. Lane’s contributors documented the engineering in meticulous detail: the corrosion chemistry of stainless steel in uranyl sulfate solutions, the handling of radiolytic gas evolution, the design of the heat exchangers, the techniques for continuous chemical processing of the fuel stream. Los Alamos had built its own variant, the Los Alamos Power Reactor Experiment, in two stages. The aqueous homogeneous approach was not a sketch. It was a working technology.
Part II, edited by H. G. MacPherson of Oak Ridge, covered molten-salt reactors. This was the technology descending directly from Wigner’s MUC-EPW-134 memorandum of May 1945 and Ray Briant’s Aircraft Reactor Experiment of November 1954. MacPherson’s section described the ARE’s successful operation in detail — a 2.5-megawatt fluoride-salt reactor that had run for 221 hours at outlet temperatures of 1,580 degrees Fahrenheit, demonstrating that molten salt fuels could be handled at temperatures no other reactor could approach. More importantly, Part II included a complete conceptual design for a civilian molten-salt power reactor, using lithium-beryllium-fluoride salts dissolving uranium tetrafluoride as fuel and thorium tetrafluoride as fertile material. The design proposed operating temperatures of 1,100 to 1,300 degrees Fahrenheit, thermal efficiencies above 40 percent, and, explicitly, “a self-contained Th-233U system with a breeding ratio of one.” The molten-salt concept, MacPherson’s section concluded, offered a plausible path to a thorium thermal breeder that could sustain itself indefinitely on domestic thorium reserves.
Part III, edited by Frank Maslan of Brookhaven National Laboratory, covered liquid-metal-fueled reactors — specifically, the Liquid Metal Fuel Reactor being developed jointly by Brookhaven and the Babcock & Wilcox Company. This was the technology Walker Cisler’s engineers had ranked first at Lake Angelus in November 1951. Brookhaven had been working on uranium-bismuth fluid fuel since 1947. A seventeen-company industrial consortium led by Babcock & Wilcox had produced a reference design in 1955. In 1956, Babcock & Wilcox had contracted with the AEC to design, build, and operate LMFR Experiment No. 1, with construction scheduled to begin in 1960. Maslan’s chapters included a full design study for a 315-megawatt-electric commercial LMFR: uranium-233 dissolved in bismuth circulating through a graphite core, with a blanket of bismuth-thorium slurry breeding new uranium-233. The reactor had a projected breeding ratio greater than one. It was designed to fuel itself, indefinitely, on domestic thorium.
Three parallel fluid-fuel reactor families. All of them capable of operating on the thorium-U233 cycle. All of them under active AEC development. All of them published openly, in detail, to the international community, at the Second Geneva Conference in September 1958.
This was the actual state of American civilian nuclear reactor technology at the moment of its maximum public flowering. The Smyth Report’s one-paragraph dismissal had not been the final word. In the decade after the war, working quietly under AEC contracts at Oak Ridge and Brookhaven and Los Alamos and Argonne, and in private partnership with Babcock & Wilcox and Westinghouse and seventeen other companies, American engineers had done exactly what Wigner had proposed in May 1945. They had built fluid-fuel reactors. They had operated them. They had designed civilian power versions of them. They had documented the technology in a level of engineering detail sufficient for the United States government to publish nine hundred pages of it, openly, to sixty-eight foreign nations.
And they had done it because the government had chosen, in the late 1940s and through the early 1950s, to fund the work. The AEC had not been monolithic in its reactor development priorities. Its naval reactor division, under Rickover, had been building pressurized-water reactors for the Navy. Its civilian reactor development division, under Lawrence Hafstad and then his successors, had been funding Oak Ridge and Brookhaven and Los Alamos to explore the fluid-fuel alternatives. Both programs were genuine. Both programs had produced operating reactors. Both were represented at Geneva 1958.
The question, as the conference closed and the delegates went home, was which of the two programs — the light-water family Rickover had built for the Navy, or the fluid-fuel family the national laboratories had developed in parallel — would define the American civilian nuclear industry over the coming decade. The utility orders that were beginning to arrive at Westinghouse and General Electric suggested an answer. Shippingport was operating. Yankee Atomic, Dresden Unit 1, and Consolidated Edison’s Indian Point were under construction. Every one of them was a solid-fueled, water-cooled reactor. None of them was a fluid-fuel reactor. The industrial infrastructure that was taking shape to support the utility orders — the fuel fabricators, the steam generator manufacturers, the engineering firms — was shaping itself around the PWR and BWR templates that Rickover’s Navy had pioneered.
But the AEC had not yet made its own formal choice. Three programs had been presented at Geneva. All three were still funded. All three had project directors, laboratories, reference designs, and schedules running into the 1960s. The Commission would have to decide, at some point, whether to continue supporting all three, concentrate on one, or let industrial drift do the choosing for it. That decision, as it turned out, would be made over the following winter.
The Task Force
In the autumn of 1958, with the Geneva Conference barely concluded and its proceedings still being bound for publication, the Atomic Energy Commission faced a quiet institutional problem. It had just told the world, in nine hundred pages, that American laboratories were developing three promising fluid-fuel reactor concepts in parallel. It had neither the budget nor the political mandate to continue all three indefinitely, as separate programs with separate project directors and separate industrial partners, through the demonstration-reactor and prototype phases that each of them required. At some point, the Commission would have to choose.
In November 1958, the Commission formed what it called the Ad Hoc Advisory Committee on Reactor Policies and Programs — a body charged, among other things, with reviewing the civilian power program and recommending which reactor concepts deserved concentrated support. The committee reported to the Commission on January 2, 1959. Its recommendation on the fluid-fuel question was brief and unambiguous.
*”The aqueous homogeneous, molten bismuth, and molten salt reactors all offer the possibility of reducing the cost of the fuel cycle, and the last two offer the possibility of high temperature operation. These three concepts for power reactors should be critically compared and work concentrated on the concept that appears the most promising.”*
The Commission acted on the recommendation within weeks. The Evaluation and Planning Branch of the Office of Civilian Reactors, working under the Division of Reactor Development, convened a fifteen-member Fluid Fuel Reactors Task Force in January 1959. Its charge was specific. It was to compare the three concepts on four criteria — present state of development, technical feasibility of breeding, potential power cost in mills per kilowatt-hour, and the research-and-development program required to reach a commercial reactor — and to provide the Commission with the information required to choose among them. The Task Force met continuously through January and February of 1959. Its report was issued as TID-8507 in February, under the signature of chairman Robert W. Ritzman of the AEC’s Division of Reactor Development.
The Task Force’s composition mattered. Its fifteen members included Robert Avery and Joseph Draley from Argonne, Jack Chernick from Brookhaven, James Evans and Edgar Hayes from Du Pont’s Atomic Energy Division, Titus LeClair from Commonwealth Edison representing the Edison Electric Institute, Wilson Cooper from the Tennessee Valley Authority, Jack McKamey from Ebasco Services, Carl Newman from United Engineers & Constructors, and Vincent Walker from Phillips Petroleum at the National Reactor Testing Station. Crucially, it also included the three project directors themselves: R. Beecher Briggs, director of the Homogeneous Reactor Project at Oak Ridge; H. G. MacPherson, director of the Molten Salt Reactor Program at Oak Ridge; and Francis T. Miles, director of the Liquid Metal Fuel Reactor Project at Brookhaven. Each of them sat on the committee that was about to rank his own program against its rivals.
The Task Force’s central finding was stated on page 2 of the report.
*”The molten salt reactor has the highest probability of achieving technical feasibility. This is largely due to the use of a solution fuel (as contrasted to a slurry fuel in the LMFR and the AHR), and the availability of a suitable container material (INOR-8).”*
INOR-8 was the nickel-molybdenum alloy — later renamed Hastelloy N — that Oak Ridge’s metallurgists had developed specifically for molten fluoride salt service. The Task Force’s judgment was that the MSR had the best fuel and the best containment material of the three candidates. Everything else followed from that.
The report quantified the finding in a table of “Relative Technical Difficulties.” The Task Force scored the three concepts on fuels, on materials for reactor internals, and on materials for external systems. Higher numbers meant greater difficulty. The scores, totaled across the three categories, were: aqueous homogeneous solution core, fifteen; aqueous homogeneous slurry core, nine; molten salt, five; liquid metal fuel, eight. The MSR had the lowest total of any concept. The aqueous homogeneous design that had consumed fifty million dollars and fifteen hundred technical man-years at Oak Ridge had the highest.
On breeding, the Task Force was cautious but optimistic. All three systems, it concluded, were “likely” to achieve “hold-own” breeder status — that is, breeding ratios slightly above one, sufficient to replace their own fissile inventory over time. Projected breeding ratios were 1.09 for the AHR, 1.05 for the LMFR, and 1.05 for the MSR. Only the AHR, in its slurry-blanket two-region form, had near-term prospects for doubling its fissile inventory on a schedule of “of the order of 15 years.” The MSR and LMFR could match that performance, but only with additional development work on internally-cooled graphite-moderated configurations whose design was less mature.
On cost, the Task Force found something that changed the nature of the choice entirely. The three concepts, when costed on a comparable basis for a 333-megawatt-electric reference plant, produced near-identical totals: 10.0 mills per kilowatt-hour for the LMFR, 10.7 for the MSR, 11.1 for the AHR in its two-region solution-slurry configuration. “The data,” the report observed drily, “do not indicate any very substantial overall difference in the relative cost potential of the three concepts.” None of the three was cheap by 1959 standards — the projected costs were “substantially higher than costs indicated in some earlier reports on fluid fuel reactor concepts,” the Task Force acknowledged. But none of the three was ruinously more expensive than the others.
The question, then, was not which fluid-fuel concept would produce the cheapest power. They would all produce similar power at similar cost. The question was which concept the AEC believed could actually be built, reliably, at reasonable engineering risk, within a development program of finite resources. On that question the Task Force was unanimous. It was the molten salt reactor.
The recommendation on program sizing was equally clear. “To carry through to a practical design of a commercial power plant,” the Task Force wrote, “each program requires research and development, design, construction, and operation of a reactor experiment and of a prototype reactor. The cost of the entire development program for any one concept would be something above one hundred million dollars. Ten years would be required. There is no indication that any one would be substantially less costly than the others.” Three phases — experimental reactor, prototype, full-scale plant — over ten years, at a hundred million dollars. That was what the Task Force, including the project directors of each of the three programs, understood to be the actual cost of getting a fluid-fuel civilian power reactor to commercial deployment.
The AEC acted on the Task Force’s ranking within two years, in the passive way that institutional decisions usually propagate through large bureaucracies. The aqueous homogeneous reactor program was terminated in 1961; HRE-2 was taken out of service, the development team was dispersed, and the fifty million dollars of prior investment became sunk cost. The Babcock & Wilcox Liquid Metal Fuel Reactor Experiment that had been scheduled to begin construction in 1960 was postponed indefinitely — the contract had been signed in 1956, the design work had continued through 1958, but the construction authorization and funding never came. Babcock & Wilcox, which had invested heavily in the LMFR program, redirected its nuclear division toward pressurized-water reactors for the commercial utility market, where the industrial momentum was now plainly visible. The LMFR experiment was never built. Brookhaven’s Francis Miles, who had chaired the LMFR project since the late 1940s and who had sat on the Task Force that recommended against his own program, presided over its wind-down.
Only the molten salt program, which the Task Force had ranked first, continued. And here the story becomes more complicated, because the Task Force’s second recommendation — that the chosen program receive “something above one hundred million dollars” and “ten years” of sustained support to build its experiment, prototype, and commercial-scale plant — was not followed. The Molten Salt Reactor Experiment would be built at Oak Ridge over 1962-1965, at a cost of roughly ten million dollars. It would operate successfully for four years. But the prototype the Task Force had identified as the essential next step — a larger demonstration reactor that would prove the commercial design — would not be built during the 1960s, would not be built during the 1970s, and had still not been built when the program was terminated in 1976.
The AEC had, in effect, done half of what its own Task Force had recommended. It had concentrated the fluid-fuel effort on the single concept identified as most promising, terminating the other two. But it had not provided that concept with the sustained hundred-million-dollar, ten-year development program the Task Force had unanimously agreed was necessary to carry it through to commercial deployment. The molten salt program survived, but at funding levels that were, year by year, too low to deliver the sequence of reactors that had been laid out in February 1959.
The Task Force’s report was filed as TID-8507 and entered the AEC’s technical documentation series. It was not classified. It was available to any engineer, policymaker, or historian who thought to look for it. Over the following two decades, as the molten salt program was systematically underfunded and eventually cancelled, TID-8507 would lie in the files as a reminder of what the AEC’s own fifteen-person working group, including the directors of each of the three competing programs, had concluded in February 1959: that molten salt was the most promising of the three, that its development would cost about a hundred million dollars over ten years, and that anything less would not carry it to commercial reality.
The Texas Talk
On October 4, 1957, the Soviet Union launched Sputnik, an artificial satellite about the size of a basketball that circled the Earth emitting a radio beep. The American reaction was panic — panic about education, about science, about the failure of the country’s technical institutions to keep pace with a Communist adversary. Within weeks, the White House, the Congress, the press, and the general public had concluded that the United States had fallen behind in some crucial way, and that dramatic measures were required.
Eisenhower understood what was happening politically and resisted it with the stubbornness of a man who had seen real disasters and knew that Sputnik was not one of them. He had appointed James Killian of MIT as his first science advisor, a position created specifically to insulate presidential decision-making from the panic. Over the following three years, as the space race accelerated and the military services competed for new weapons programs, Eisenhower repeatedly intervened to slow what he saw as an increasingly mindless expansion of defense spending. He was not always successful. The Strategic Air Command continued to grow. The Navy’s Polaris missile program moved forward with Rickover’s support and funding that Eisenhower had trouble restraining. The Air Force’s ICBM programs consumed billions of dollars at a pace the President found alarming.
What Eisenhower was witnessing, and trying to name, was the emergence of a permanent institutional alignment between the armed services and the industrial firms that supplied them. He had served as supreme commander of Allied forces in Europe during the largest war in history. He knew what it looked like when a nation mobilized its industrial capacity for military purposes. What he was seeing in the late 1950s was something new: mobilization without the war. A permanent wartime industrial structure, operating in peacetime, consuming a sustained share of national resources, generating its own political constituency that advocated for its continued expansion.
The Atomic Energy Commission, though civilian in its legal structure, was part of this emerging complex. Its largest budget category by a substantial margin was weapons production. The fourteen production reactors at Hanford and Savannah River; the gaseous-diffusion plants at Oak Ridge and Paducah and Portsmouth; the weapons-assembly facilities at Rocky Flats and Amarillo; the test sites in Nevada and the Marshall Islands; the laboratories at Los Alamos and Sandia and Livermore — all of this consumed budgets that dwarfed the civilian nuclear power program even at its Geneva-1958 peak. The AEC’s relationship with the Department of Defense — through the Joint Committee on Atomic Energy, through the Military Liaison Committee, through personnel like Rickover who operated simultaneously in both organizations — was intimate and consequential.
By 1958, as the fluid-fuel programs were being openly documented at Geneva, the internal AEC knowledge that plutonium was going to be in surplus was already six years old. Gordon Dean’s 1952 memorandum had predicted the surplus would arrive “by the mid- or late 1960s.” Events were proving him right. Hanford’s older reactors were aging, some approaching the end of their useful lives, and the question of whether to build replacements was already being asked internally. The weapons-production expansion of 1950-1955 had succeeded so completely that it was beginning to overshoot its own rationale.
But the overshoot did not produce retrenchment. It produced continued production. Tritium’s 12.3-year half-life ensured a permanent minimum demand for heavy-water reactors at Savannah River, even as plutonium accumulated. The industrial infrastructure that had grown around the expansion could not be easily unwound. The political constituency that the weapons complex had created — the unions, the state delegations in Washington and South Carolina and Idaho, the contractors, the scientific and engineering professional societies, the university programs that trained the workforce — had its own momentum. By the end of the 1950s, the AEC was what the weapons complex required it to be, and was continuing to become more so.
In 1960, his last full year in office, Eisenhower began working on what he intended to be his valedictory address to the American people. He worked on it with his brother Milton, who was a university president and a trusted advisor. He worked on it with his chief speechwriter Malcolm Moos, who would remain associated with the speech’s phrasing for the rest of his life. The drafts went through many revisions. The president wanted to name what he had seen. He also wanted to be careful not to be misread as an opponent of national defense — he was, after all, the author of a robust defense posture and would not disavow it — but as someone who understood the difference between a defense establishment that served the nation and one that had developed its own momentum.
The phrase “military-industrial complex” appeared in late drafts. It was not the first such phrase that had been considered. Earlier drafts had used “military-industrial-congressional complex,” which was more accurate historically but which the president felt was politically impolitic. The final version named only two of the three institutional partners, leaving Congress out of the direct indictment. But the meaning was clear enough.
The Farewell
On the evening of January 17, 1961, three days before John F. Kennedy’s inauguration, Eisenhower sat at his desk in the Oval Office and delivered his farewell address to the American people. The speech was short, not quite ten minutes. It was broadcast live on television. The cameras stayed on his face throughout.
He began conventionally enough. He thanked the American people. He wished his successor well. He reviewed the accomplishments of his administration. And then, midway through the speech, he turned to the warning he had been working on for over a year.
He described the change in America’s defense posture. Before the Second World War, he said, the United States had no permanent armaments industry. American manufacturers of plowshares could, “with time and as required, make swords as well.” But the postwar world no longer permitted the luxury of emergency improvisation. The country had been compelled to create a permanent armaments industry of vast proportions. Three and a half million men and women were now directly engaged in the defense establishment. Annual spending on military security exceeded the net income of all American corporations combined.
“This conjunction of an immense military establishment and a large arms industry is new in the American experience,” he said. “The total influence — economic, political, even spiritual — is felt in every city, every statehouse, every office of the federal government. We recognize the imperative need for this development. Yet we must not fail to comprehend its grave implications. Our toil, resources, and livelihood are all involved; so is the very structure of our society.”
Then the phrase itself: “In the councils of government, we must guard against the acquisition of unwarranted influence, whether sought or unsought, by the military-industrial complex. The potential for the disastrous rise of misplaced power exists and will persist.”
He went further. He warned about the growing relationship between the federal government and the nation’s scientific establishment. The “prospect of domination of the nation’s scholars by federal employment, project allocations, and the power of money is ever present,” he said, “and is gravely to be regarded.” In a passage that would matter particularly for the nuclear story, he added: “Yet, in holding scientific research and discovery in respect, as we should, we must also be alert to the equal and opposite danger that public policy could itself become the captive of a scientific-technological elite.”
What Eisenhower was describing, though he did not name it specifically, was exactly the institutional structure that had shaped American civilian nuclear power over the preceding fifteen years. The AEC had been a creature of the weapons program in its origins and had remained one in its institutional orientation. The expansion of 1950-1955 had written the weapons priorities into the agency’s DNA. The civilian power reactors being built in the late 1950s and planned for the 1960s were descendants of naval technology — Rickover’s submarine work, transplanted to utility grids — rather than descendants of the fluid-fuel research that the AEC had also funded and had openly disclosed at Geneva only twenty-eight months before the farewell speech was given. The reactors Cisler’s engineers had identified as superior in 1951 had not been built, and with each passing year were less likely to be built, because the military-industrial complex did not require them. The reactors that were being built were the reactors that structure required.
Eisenhower did not say any of this explicitly. He did not mention thorium. He did not mention Weinberg. He did not mention Oak Ridge or Shippingport or the Molten Salt Reactor Experiment that was then being planned. He did not mention Geneva 1958, though his administration had presided over it. What he named was the larger pattern: the drift of national decision-making into the hands of institutions whose interests were served by continued expansion of the defense-industrial enterprise. The thorium story was one particular consequence of that pattern, but it was not, to the outgoing president, any more visible than many other consequences that were invisible to him.
He closed the speech with a prayer for peace. He asked the American people to be vigilant. Three days later, John F. Kennedy was sworn in as the thirty-fifth president of the United States.
Epilogue
In January 1961, the American civilian nuclear industry was fifteen years old. Shippingport had been operating for three years. The Yankee Atomic plant in Massachusetts was generating power. Dresden Unit 1, a General Electric boiling-water reactor near Chicago, was approaching completion. Cisler’s Fermi-1, the fast breeder at Lagoona Beach, was in the middle of its long and troubled construction. The AEC’s Power Reactor Demonstration Program had solicited proposals from the industry and received, among other bids, the proposal from the Power Reactor Development Company for the plant that would become Fermi.
Weinberg, at Oak Ridge, was planning the Molten Salt Reactor Experiment. Construction would begin in 1962. It would reach criticality in 1965.
The aqueous homogeneous reactor program that had occupied the first section of *Fluid Fuel Reactors* was nearing formal termination, its fate effectively decided by the Task Force’s ranking two years earlier and by a series of corrosion incidents at HRE-2 that made restart impractical. The Brookhaven / Babcock & Wilcox Liquid Metal Fuel Reactor Experiment that had been scheduled for construction in 1960 had not been built. It would never be built. Babcock & Wilcox, which had held the contract since 1956, redirected its nuclear division toward pressurized-water reactors for the growing utility market, where the money was. The Task Force’s recommendation that work be “concentrated on the concept that appears the most promising” had been implemented for the two losers. Whether it would be fully implemented for the winner — whether the molten salt program would receive the hundred million dollars and ten years the Task Force had said were necessary — remained an open question as the new administration took office.
Rickover, at Naval Reactors, was building nuclear-powered carriers and cruisers in addition to submarines. The *Enterprise*, the Navy’s first nuclear-powered aircraft carrier, would be commissioned later in 1961. Rickover would not retire for another twenty-one years, by which time he would be the longest-serving active-duty officer in American history.
The weapons complex continued. The fourteen production reactors at Hanford and Savannah River continued. Dean’s 1952 prediction that plutonium would reach surplus “by the mid- or late 1960s” was on schedule to prove correct. The strategic logic that had built the complex was approaching its own obsolescence. The institutional structure built on that logic was not.
Seaborg, who had spent the 1950s as associate director of the Lawrence Berkeley Laboratory while continuing his work on transuranium elements, had been visited in late 1960 by John Kennedy, who wanted him to chair the Atomic Energy Commission in the new administration. Seaborg accepted. He would serve for ten years, through three presidents, and would be in the control room at Oak Ridge on August 8, 1967, when the Molten Salt Reactor Experiment achieved criticality on uranium-233 for the first time — closing, personally, the circle he had opened on April 23, 1941.
Eisenhower retired to a farm in Gettysburg, Pennsylvania. He lived another eight years. He never spoke publicly about the specific technologies his administration had funded or failed to fund. He had said what he had to say in the farewell address.
What he had named, the institutional structure that shaped American nuclear power in the first era, would shape it for the rest of the twentieth century. Of the three fluid-fuel reactor families documented at Geneva 1958 — aqueous homogeneous, molten-salt, liquid-metal-fueled — the first would be formally terminated within months of Eisenhower’s farewell, the third had effectively been terminated in the year before it, and only the second would survive long enough to achieve its own demonstration-reactor criticality four years later. That reactor, the MSRE, would be built, would succeed, and would be cancelled. Fermi-1 would fail in ways that damaged the entire idea of nuclear power. The pressurized-water reactors that Shippingport had prefigured would multiply through the 1960s and early 1970s, would stop being ordered after Three Mile Island and the cost overruns that preceded it, and would continue operating into the twenty-first century as the backbone of American civilian nuclear generation.
The thorium story, meanwhile, would wait. It would wait in the classified files that had accumulated since 1944. It would wait in the ORNL reports that Weinberg and his colleagues would keep producing into the 1970s. It would wait in the nine hundred pages of *Fluid Fuel Reactors* that had gone out of print and were not reprinted. It would wait through the decades during which American nuclear power was defined by the technology that Rickover had built for the Navy and that the industry had inherited.
Notes on Sources
Part Two draws primarily on the AEC official history volumes, particularly *Atomic Shield, 1947–1952* by Richard Hewlett and Francis Duncan (Pennsylvania State University Press, 1969) and *Atoms for Peace and War, 1953–1961: Eisenhower and the Atomic Energy Commission* by Richard Hewlett and Jack Holl (University of California Press, 1989). Francis Duncan’s *Nuclear Navy, 1946–1962* (University of Chicago Press, 1974) is the indispensable source for Rickover’s role and for the Bettis laboratory’s work on the pressurized-water reactor that became Shippingport. David Lilienthal’s *Atomic Energy: A New Start* (Harper & Row, 1980) and his earlier published journals provide his own account of his years as AEC chairman.
The account of the Smyth Report in the Prologue draws directly from Henry DeWolf Smyth, *Atomic Energy for Military Purposes: The Official Report on the Development of the Atomic Bomb under the Auspices of the United States Government, 1940-1945* (Princeton University Press, 1945), particularly paragraphs 2.21-2.25. The characterization of what the Smyth Report omitted relies on the Met Lab meeting minutes MUC-LAO-17, MUC-LAO-18, and MUC-LAO-30 from April 1944 and on MUC-EPW-134 from May 17, 1945; the Seaborg journal entries for April 23, 1941 and the December 21, 1944 memorandum on U-232 contamination pathways are discussed in Part One of this book.
The account of the Savannah River Site and the 1950-1955 weapons-production expansion draws extensively on Rodney P. Carlisle and Joan M. Zenzen, *Supplying the Nuclear Arsenal: American Production Reactors, 1942-1992* (Johns Hopkins University Press, 1996), prepared under a Department of Energy contract by History Associates Incorporated. Carlisle and Zenzen’s detailed account of the five Savannah River heavy-water reactors, the three additional Hanford graphite reactors, and the Gordon Dean 1952 memorandum predicting plutonium surplus by the mid-to-late 1960s provided the factual basis for Chapter Seventeen.
Chapter Fifteen, on William L. Borden, draws from Borden’s own book, *There Will Be No Time: The Revolution in Strategy* (New York, 1946), particularly the foreword (Borden’s account of the November 1944 V-2 sighting), Chapter Two (“The Certainty of War Amidst Anarchy”), and Chapter Four (“The Pattern of Atomic Warfare”), which contain the core thesis that peacetime stockpile expansion was the only rational response to the atomic age. Borden’s subsequent career as executive director of the Joint Committee on Atomic Energy from 1949 to May 1953 is documented throughout Hewlett and Duncan’s *Atomic Shield*.
Chapter Seventeen, on the Arsenal Rises, draws from three principal sources beyond Carlisle and Zenzen. Hewlett and Duncan’s *Atomic Shield, 1947-1952* (Pennsylvania State University Press, 1969) provides the institutional narrative of the 1949-1952 weapons-complex expansion, including the McMahon-Borden letter of July 14, 1949, Borden’s fall 1949 assessment of Soviet production capability, the October 1949 GAC recommendation and its political aftermath, McMahon’s November 1949 letter to Truman, and the November 1951 Military Liaison Committee discussion in which Dean responded to the suggestion of obtaining thorium ore by noting that there were no plans to develop weapons using uranium-233. The specific quotation from McMahon’s December 1950 phone call to Dean (“I know you are fully conscious of living every day in a Pearl Harbor atmosphere…”) and other McMahon-Dean and Borden-Dean exchanges are drawn from Roger M. Anders, ed., *Forging the Atomic Shield: Excerpts from the Office Diary of Gordon E. Dean* (University of North Carolina Press, 1987). Anders’s editorial introductions to each chapter synthesize the institutional context; the diary entries themselves provide the dated record of specific phone conversations between Dean and the Joint Committee leadership. The account of Dean’s law partnership with McMahon beginning in 1940 is from Anders’s chronology and editorial essays. McMahon’s September 18, 1951 Senate speech advocating an “atomic army, navy, and air force,” and McMahon’s July 1952 deathbed telephone pledge to the Connecticut Democratic convention to produce hydrogen bombs “by the thousands,” are both documented in Hewlett and Duncan.
Walker Cisler’s autobiography *A Measurable Difference* (Wayne State University Press, 1983) and the *Detroit Edison* corporate history supply the Detroit Edison account of the Lake Angelus meeting and the early PRDC organization. The detailed account of the Dow-Detroit Edison study group’s reactor ranking is drawn from the Fermi-1 technical history (ANS, 1979).
The account of Geneva 1958 and the three fluid-fuel reactor programs at their moment of maximum public visibility draws on James A. Lane, H. G. MacPherson, and Frank Maslan, eds., *Fluid Fuel Reactors* (Addison-Wesley, 1958), published as a contribution to the Second International Conference on the Peaceful Uses of Atomic Energy, Geneva, September 1958. The three parts of that volume — edited by Lane for aqueous homogeneous reactors (Oak Ridge), MacPherson for molten-salt reactors (Oak Ridge), and Maslan for liquid-metal-fueled reactors (Brookhaven) — provided the technical and institutional basis for Chapter Twenty-Three. The specific details of the Brookhaven/Babcock & Wilcox LMFR program, including the 1955 seventeen-company consortium study and the 1956 AEC contract for LMFR Experiment No. 1, are drawn from Maslan’s preface and Chapter 24 of that volume.
Chapter Twenty, on the Aircraft Reactor Experiment, draws on three 1957 papers in *Nuclear Science and Engineering* authored by the Oak Ridge team that built and operated the ARE: E. S. Bettis et al., “The Aircraft Reactor Experiment — Design and Construction,” *NSE* 2:804-825 (1957); W. K. Ergen et al., “The Aircraft Reactor Experiment — Physics,” *NSE* 2:826-840 (1957); and E. S. Bettis et al., “The Aircraft Reactor Experiment — Operation,” *NSE* 2:841-853 (1957). These papers provide the definitive technical record of the ARE, including the design evolution from solid-fuel through stagnant-liquid to circulating-liquid configurations, the fuel salt chemistry (NaF-ZrF₄-UF₄ at 93.4% ²³⁵U), the operational milestones (first criticality 3:45 PM November 3, 1954; peak power 2.5 MWt; 221 hours of nuclear operation; 96 MWh total energy production), and the key physics findings (the strongly negative temperature coefficient of –6 × 10⁻⁵ Δk/k per °F and the absence of xenon-135 poisoning due to continuous fission gas removal at the pump bowl). The Aircraft Nuclear Propulsion program context — NEPA at Fairchild, the 1951 ORNL reorganization, Briant’s project directorship — is from Richard G. Hewlett and Francis Duncan, *Atomic Shield, 1947-1952*, supplemented by H. G. MacPherson’s retrospective on the MSR program origins.
Chapter Twenty-One, on the Mark I prototype and the USS *Nautilus*, is drawn almost entirely from Richard G. Hewlett and Francis Duncan, *Nuclear Navy, 1946-1962* (University of Chicago Press, 1974), particularly Chapter 6 (“Prototypes and Submarines”) and Chapter 7 (“Toward a Nuclear Fleet”). Duncan, the Navy’s reactor historian at the time of writing, had direct access to the Mark I project records, the Electric Boat construction files, and personal interviews with Rickover, Kintner, Wilkinson, and the senior Westinghouse engineers. The specific narrative of the 100-hour run in the summer of 1953, including Rickover’s decision to post a chart of the North Atlantic and his refusal to shut down as steam-plant components began to act erratically at the sixtieth hour, is drawn from E. E. Kintner, “The First Days of the Mark I,” *Journal of the American Society of Naval Engineers* 72 (February 1960): 9-13, and Kintner, “Admiral Rickover’s Gamble,” *The Atlantic Monthly* 203 (January 1959): 31-35, both quoted in Hewlett and Duncan. The Mark I’s first criticality time (11:17 PM March 30, 1953) is from the reactor’s original log book at the Naval Reactor Facility. The account of the January 17, 1955 Nautilus departure — including the loose locking pin on the starboard reduction gear and the “UNDERWAY ON NUCLEAR POWER” signal to the Skylark — is from the contemporaneous reports of Rickover to the CNO dated January 19, 1955, and of Wilkinson on the trials. The Atlantic Fleet exercise assessment of August 1955, including Admiral Jerauld Wright’s conclusion that “no future combatant submarine be built that is not nuclear powered,” is from CINCLANTFLT to CNO, October 11, 1955, as cited in Hewlett and Duncan.
Chapter Twenty-Four, on the Fluid Fuel Reactors Task Force, draws directly from the Task Force’s own report: *Report of the Fluid Fuel Reactors Task Force to the Division of Reactor Development, United States Atomic Energy Commission*, TID-8507, February 1959. The report was prepared by a fifteen-member task force convened by the Evaluation and Planning Branch of the Office of Civilian Reactors in response to a January 2, 1959 recommendation from the Ad Hoc Advisory Committee on Reactor Policies and Programs. The specific findings quoted in the chapter — the molten salt reactor’s designation as having “the highest probability of achieving technical feasibility,” the relative technical difficulty scores, the projected breeding ratios, the power cost comparison, and the hundred-million-dollar, ten-year program estimates — are all drawn from sections II through VIII of TID-8507. The chapter’s account of the Task Force’s composition, including the participation of the three project directors Briggs, MacPherson, and Miles, is drawn from the report’s preface and membership list.
Weinberg’s 1997 speech to a Korean scientific delegation provides his most condensed retrospective account of the Aircraft Reactor Experiment and the origins of the molten-salt concept at Oak Ridge.
The account of Eisenhower’s farewell address relies on the standard published text of the speech, available through the National Archives and the Eisenhower Presidential Library. The context of the speech’s drafting, including the progression through multiple versions and the role of Milton Eisenhower and Malcolm Moos, is well documented in the Eisenhower Library’s online archive and in the Hewlett and Holl volume.