The decision to assemble Manhattan Project scientists at Los Alamos, followed less than three weeks later by the chain reaction under Stagg Field, gave impetus and clarity to the project. There were many technical problems to be solved and strategic decisions to be made, but what Groves and Oppenheimer knew they needed, as soon as possible, was a fissionable core for the bomb. It could be made of U-235 or plutonium (Pu-239), with the precise amount of these materials needed remaining a matter of speculation, though not wild speculation. To produce both substances the project would need as much U-238 ore as Groves could put his hands on. Here was a task Groves readily understood, and he undertook it with his usual relentless determination. He believed, at first, that monopolizing the world’s uranium supply was possible. The Germans had Joachimsthal, but the United States had Sengier, who not only sold his Staten Island supply to Kenneth Nichols but who promised another 3,000 tons from the Congo. (In the end, the United States would amass some 6,000 tons of uranium during the war. The Congo was the source for 3,700 tons, Canada’s Great Bear Lake 1,100, and the rest came from the United States itself.)
Groves also hoped to control the world’s supply of thorium, a radioactive element often contained in monazite sands, which were abundant in the Netherlands East Indies, Brazil, and especially the Travancore Coast of southern India. In all these gathering efforts he gained the cooperation of the British.36
Having secured his uranium, Groves now faced the need to refine it on a mass scale. Compton and others had planned to erect a pilot plant for making plutonium in the Argonne Forest, 20 miles outside Chicago. It emerged, however, that the uncertainties of plutonium production made risky the presence of a plant, even the so-called semi-works, so near a big city, and it seemed more sensible to place it closer to the full-scale facility where the production work would actually occur. Groves, it will be remembered, had already purchased a large plot of land in the Tennessee Valley, as close to modest Knoxville as Argonne was to Chicago. But the immediate area was sparsely populated and its residents poor. Through the winter of 1942-3 contractors descended on the place, building a railroad extension, laying down new roads, and putting up homes for workers and plant facilities staggering in their size and facelessness: the enormous dark box of the K-25 uranium separation plant covered some 42 acres. The town thus created was called Oak Ridge (its inhabitants dubbed it ‘Dogpatch’ after the rundown spot in the L’il Abner comic strip); the place altogether was known as the Clinton Engineering works. To the site came thousands of workers. Many were not sure what they were supposed to be making. Those who thought they knew were compartmentalized and sworn to secrecy. Mail was censored, phone calls monitored, and when boredom set in there was little for entertainment except movies and games of checkers.37
They were trying at Oak Ridge to make quantities of U-235. By the summer of 1943 scientists at Los Alamos were calculating that they would need some 88 pounds (40 kilograms) of U-235 to build the kind of bomb they had in mind. Part of the Oak Ridge enterprise was given over to Ernest Lawrence’s electromagnetic separation technique, of which both he and Conant were enamored. They wanted Groves to build
2,000 Calutrons there. Groves, less convinced but nevertheless willing to place at least some of his chips on the magnets, built 500. In theory, these Calutrons should have produced enough U-235 for a bomb within two years, but design and construction problems resulted in the great machines’ shorting out with dismaying frequency. Mice and birds found their way into the Calutrons and shut them down until the animals’ remains were discovered and removed. By late 1943 the Calutrons had yielded virtually nothing. Much of the rest of the Clinton Works was devoted to separation by gaseous diffusion, in which Harold Urey played a key role. Yet here, too, manifold problems existed, especially with the manufacture of the delicate metallic barriers that were to filter the uranium hexafluoride gas through a series of cascades, producing the U-235 isotope. Debate over the composition of the barriers raged into early 1944, theory foundering more than once as it encountered the realities of engineering.38
Groves and the scientists had also originally intended to make plutonium at Oak Ridge. But the Tennessee Plant could not do everything, and if something went wrong with the plutonium-making process, and ‘the wind was blowing through Knoxville’, as Groves worried it might, there could be substantial loss of life, a shutdown of the Calutrons and gas diffusers, and, worst of all, a breach of security. Groves wanted yet another site on which to build a plant to conjure plutonium. His criteria were water power, a favorable climate, and, above all, isolation—‘at least twenty miles between the piles and separation area and the nearest existing community of one thousand or more inhabitants’. John Dudley had helped Groves find Los Alamos; in December 1942 the general asked Lieutenant Colonel Franklin T. Matthias to locate a place to make plutonium. With two engineers from the Du Pont Company, Matthias settled on a high desert cut by the Columbia River in southern Washington State, near the small town of Hanford, population 100. Groves reviewed the site and approved.39
Another massive building project ensued. Construction crews were recruited to live in barracks, segregated by sex and race, paid somewhat higher than wartime scale, and treated to an abundance of good food— a treat in 1943. They were building a city, one with the single purpose of producing an elusive element for a mysterious project based elsewhere. The Hanford site proved wilder than Oak Ridge. At one point there were over 50,000 people working in the remote desert, putting in nine hours daily and extra time nights and Sundays. They entertained themselves as best they could, in an enormous beer hall, a gambling hall with slot machines, movie theaters, and a bowling alley. There were fights—‘occasionally bodies were found in garbage cans the next morning’, wrote the physicist John Marshall—suicides, and prostitution.40
Du Pont was responsible for design and construction at Hanford; the Manhattan Project thus involved not only Big Science but Big Business. Some scientists were as suspicious of the men in suits as they were of those in khaki uniforms. Groves’s position, shared by Bush, Compton, and Lawrence, was that no one but a large firm could do a job on this scale. Needed at Hanford were three reactors, exalted versions of Fermi’s pile, and four plants at which plutonium would be chemically separated from slugs of uranium that had undergone a chain reaction within the piles. Eugene Wigner designed the reactors: enormous cylinders of pure graphite shot throughout with aluminum tubes, into which went 200 tons of uranium metal slugs, themselves ‘canned’ in aluminum. Water from the Columbia River coursed through the tubes and around the cans, cooling the uranium as it reacted. Cooked in the reactor for at least 100 days, the canned slugs were pushed out of the back of the pile into pools to contain their radioactivity, and months later—two months were the minimum necessary for safety, with four more desirable—they were taken off to the separation plants to have their bits of plutonium teased out. From the start there was a serious glitch. The Hanford B reactor, run at full power as no test reactor had been previously, produced quantities of the element xenon, which absorbed neutrons and ‘poisoned’ the chain reaction. The engineers and scientists determined to overmatch the poisoning by stuffing more uranium slugs into extra tubes Du Pont had drilled into the graphite. The B reactor was restarted.41
In the summer of 1944 Oppenheimer recommended to Groves the preenrichment, by thermal (hot to cold plate) diffusion, of feed uranium for the Oak Ridge Calutrons. That made production of U-235 creep toward the level needed for a bomb—64 kilograms in the event. At Hanford, progress was steady once the xenon poisoning problem was solved, but still too slow for Groves’s taste. He ordered Du Pont officials to move things along: he needed roughly 6 kilograms for a test shot and another 6 for the first plutonium bomb. Du Pont obliged, with Groves’s permission, by taking shortcuts, among them reducing the amount of time workers left the radioactive slugs in their post-reactor baths. That greatly increased the danger to those who then transported the slugs to the separation plants, and especially to those who then removed the slugs from the aluminum cans, which meant dissolving the aluminum in acid. Groves decided he could live with the risk, and that his workers could too, especially if they were not informed of its possible magnitude. Los Alamos got its first delivery of Hanford plutonium in February 1945. Taken by convoy, escorted by men wielding shotguns and submachine guns, the stuff came in stainless steel flasks, each holding, writes Robert Norris, ‘eighty grams of the bluish green slurry’.42