The Apocalypse Factory

by Steve Olson

  In 1943 Woods married John Marshall, a young physicist who had moved with Fermi from Columbia to work at the Met Lab. After Fermi’s successful demonstration of Chicago Pile 1, the reactor was taken apart and moved to a laboratory about 20 miles west of Chicago, where it was renamed Chicago Pile 2. At the new lab, Woods was helping Fermi measure the rate at which various materials absorbed neutrons from the reactor, a job she thought she might lose when she became pregnant not long after getting married. She decided to hide the pregnancy beneath her baggy overalls, already bulging with a tape measure, slide rule, and notebooks. Her fellow scientists never learned that she was pregnant. But Fermi knew and was worried that he would have to deliver the baby if it arrived early, so he asked his wife for instructions on what to do. Woods later recalled:

  When he told me he was ready, it stiffened my resolution that under no circumstance would he get the chance to practice midwifery, which, in retrospect, was no doubt a disappointment to him. Anyway, the question became academic. I went to the hospital a couple of days early with high blood pressure, and came out with a baby and was back at work in a week, not quite on a par with Italian peasants, but close to it.

  When Woods and Marshall moved to Hanford in 1944, their job was to “babysit the reactors,” as she later described it. She was the only woman scientist there—the construction workers built her a private bathroom in the reactor building where she worked. But she could not begin her babysitting until the reactors were up and running, and by the time she arrived construction was well behind schedule. Part of the problem had been getting enough men to build the facilities. Initially, Hanford workers had their hands full just building the construction camp. They also had hundreds of miles of railroads, water and sewer pipes, and electrical lines to lay before they could even begin the production plants.

  An even bigger reason for the delay was getting the facilities designed. Over the course of 1943, the engineers at DuPont and the scientists at the Metallurgical Laboratory settled into an uneasy working relationship. The Met Lab scientists, in consultation with the DuPont engineers, were in charge of the basic design of the production plants. But the detailed designs, including blueprints, came largely from Wilmington. The blueprints then were reviewed for accuracy by the Met Lab scientists, who grumbled about having to check engineers’ work. The two camps argued incessantly.

  Crawford Greenewalt was in charge of coordinating the work being done in Chicago and Delaware. The Met Lab scientists, he later said,

  suffered from a general disease that brilliant people, particularly in physics, seem to have; that is, because they’re brilliant in their own field, they think they know everybody else’s. Wigner [the leader of the reactor design team in Chicago] would have had not the slightest hesitation in telling us how to run the DuPont Company. As a matter of fact, all of the difficulty—and there was a great deal during that design—was their certainty that they knew better than we how to go at this problem, . . . as if [engineers] were glorified plumbers that had no science of their own.

  The addition of Groves and the military to the mix heightened tensions. The scientists could already feel themselves forfeiting control to the DuPont engineers, whom they accused of wanting to corner the market in nuclear energy after the war. Now they could see a worse prospect on the horizon—a military takeover of all matters nuclear. A particular flashpoint for the scientists was the practice known as compartmentalization. As Groves described it, compartmentalization was at the heart of security. “My rule was simple and not capable of misinterpretation—each man should know everything he needed to know to do his job and nothing else.” If someone expressed too much interest in what someone else was doing, that person was immediately under suspicion. The only one who knew what everyone was doing—and therefore the only person in a position to make informed decisions about the project as a whole—was Groves.

  The secrecy inherent to compartmentalization was anathema to the Met Lab scientists. They were used to working as they did in academia—as members of open and overlapping committees. They needed to know what other people were doing to make progress, because an idea that they needed could come from anywhere. Asking them to “stick to their knitting,” as Groves once described his intentions, was asking them to shackle themselves.

  On this and many other issues, Groves and Szilard, who remained at the Met Lab throughout World War II, immediately locked horns. Szilard took special pleasure in “baiting brass hats,” and Groves was an easy target. After one of Groves’s typically overbearing presentations to the Chicago scientists, Szilard turned to his colleagues and said, “You see what I told you? How can you work with people like that?” For his part, Groves, contemplating Szilard’s heavy German accent, nebulous responsibilities, and alien enemy status, suspected he was a spy. Within weeks of meeting him, Groves drafted a letter to be signed by Secretary of War Stimson ordering that Szilard be interned for the duration of the war. Stimson refused to sign, and when Groves tried to have him fired, Szilard appealed to Compton and the other scientists in Chicago to keep his job. Groves retaliated by putting Szilard under surveillance for the duration of the war. It was a comical arrangement. Szilard would sometimes lead his tails on chaotic walks through the cities he visited. Other times, he took pity on the agents and invited them to join him for a taxi ride or cup of coffee.

  In other circumstances, the three-way partnership among the Met Lab scientists, the DuPont engineers, and the Corps of Engineers would never have worked. The groups were too antagonistic, too divergent in their outlooks. But all three wanted the same thing—atomic bombs as quickly as possible. As Matthias said of the arrangement, “We were a team, determined to win.”

  FINALLY, BY OCTOBER 1943, the conceptual designs were far enough along to begin digging the foundation for Hanford’s first reactor. Surveyors had laid out six reactor sites, labeled A through F, strung along the bank of the Columbia River, though in the end only three reactors were needed to produce the plutonium for the Manhattan Project. The first would be built on the B site; since then, it has been called the B Reactor. The D Reactor and F Reactor would follow in quick succession.

  Once the ground had been excavated and the forms set, concrete workers began pouring a foundation 23 feet thick atop the sand, gravel, cobbles, and boulders deposited over eons by the Columbia River. African American workers specialized in the pouring of concrete. “We wore rubber boots, hard hats, slicker pants, gloves to keep the concrete from messing your clothes,” said Willie Daniels, a construction worker from Oklahoma who arrived at Hanford with his brother in the summer of 1943. “I knew what I was doing when it came to spreading mud.” As the walls of the building rose, workers pumped the concrete through steel pipes. Their first day at work, Daniels and his brother made $19.20 between the two of them—more than his brother brought home in a month at his former railroad job.

  From late 1943 to early 1945, the Manhattan Project built three nuclear reactors along the right bank of the Columbia River. Two separation plants on a plateau in the center of the Hanford reservation separated plutonium from the irradiated fuel elements. On this map the outlines of structures are drawn larger than their actual sizes. The distance from the B Reactor to the T Plant is about five miles. Map courtesy of Matt Stevenson; drawing courtesy of Sarah Olson.

  Atop the foundation was laid a steel base plate and then a layer of ten-inch-thick cast-iron blocks cooled by water pipes. These blocks were so carefully machined that their top face was level to within one two-hundredth of an inch—about the thickness of a plastic shopping bag. Along the edges of the base plate rose more cast-iron blocks, forming walls for the reactor. But these would not be nearly enough to protect people from the reactor’s intense radioactivity. For that, the designers came up with an ingenious shield. It consisted of four feet of alternating steel and Masonite layers. The steel would absorb slow neutrons and gamma rays. The hydrogen atoms in the Masonite, which is a type of pressure-cooked hardboard, would slow down the
fast neutrons from the reactor, which then would be captured in the steel. Together, the 4 feet of steel and Masonite were as effective as 15 feet of concrete.

  By May 20, 1944, enough of the walls and roof were done to begin building the reactor itself. About a dozen construction workers at any one time were dressed in white smocks and booties, all carefully laundered in boron-free soap to keep the neutron-swallowing element out of the reactor. They began to lay down more than 75,000 graphite blocks—direct descendants of the graphite blocks Fermi and Szilard had used first in New York and then in Chicago. About four inches square and 48 inches long, each of the high-density blocks weighed 50 to 60 pounds—milling them was “about as tough as milling iron,” said one of the machinists. Workers had to keep the blocks meticulously clean. Oil from the planers or drills, or even the workers’ own sweat, would absorb neutrons and slow down the chain reaction. After each layer of blocks was laid, workers carefully vacuumed the graphite to make sure no contaminants remained. Layer by layer, a mass of graphite the size of a two-story house took shape inside the B Reactor.

  Down the long axis of about one-fifth of the graphite blocks, the machinists had drilled a hole two inches across. As the blocks were stacked in the reactor, these holes lined up like gun barrels extending from the front of the reactor to the back. Into each of these holes, the reactor’s builders inserted a long aluminum tube. Like the graphite blocks, the aluminum tubes had to be kept clean during the construction and operation of the reactor. “They had carloads of Kotex coming in—I mean a lot of Kotex, which made a lot of people wonder why,” recalled one construction worker. But sanitary pads were the perfect size to keep the process tubes clean. “That’s the way we swabbed all those pipes.”

  The design developed in Chicago and Wilmington called for workers to insert fuel elements containing uranium into each of these tubes. The uranium-235 atoms in the uranium would start fissioning, just as they had in Fermi’s Chicago reactor, and the neutrons from the fissions would both keep the chain reaction going and convert uranium-238 atoms to uranium-239. After a few weeks of operation, about one of every 10,000 uranium-238 atoms would absorb a neutron. Operators would then shove fresh uranium fuel elements into the process tubes, which would cause the cooked fuel to drop out of the back of the reactor into a pool of water. After a couple of weeks to allow the uranium-239 to decay into neptunium-239 and then plutonium-239, the fuel elements would go to the separation plants to have plutonium-239 extracted from the uranium.

  The uranium fuel elements would generate immense amounts of heat inside the reactor. To keep the fuel from melting, the fuel elements would take up most but not all the space within the aluminum tubes. The rest of the space would be filled with Columbia River water pumped from the front of the reactor to the back. The hydrogen atoms in the water would absorb neutrons, but it would be moving through such a confined space that each aluminum tube would hold only about a gallon of water at any given moment. The designers calculated that water would enter the reactor at about 50 degrees Fahrenheit (the Columbia is a cold river for swimming). About two seconds and 40 feet later, it would exit the reactor too hot to touch.

  In addition to the holes for the process tubes, the graphite blocks had holes drilled crosswise to accommodate two sets of control rods, like the ones Fermi used in Chicago to control his first reactor. Nine water-cooled rods extended into the left-hand side of the reactor, each of which was covered with boron to absorb neutrons. While the reactor was being loaded with uranium, the inserted control rods would keep a chain reaction from occurring. Operators would then withdraw the control rods to fire up the reactor. In addition, 27 safety control rods hung above the reactor like swords ready to plunge into a stone. If the reactor seemed to be getting out of control, these rods would drop into vertical holes drilled in the graphite blocks. The reactor, when it was completed, was like a giant Jenga puzzle riddled by holes in all three directions.

  A crucial design decision involved the number of process tubes in the reactor. The design from the Chicago scientists called for about 1,500 process tubes arranged in a roughly circular pattern. According to their calculations, that should be plenty to operate the reactor. But the DuPont engineers, and at least some of the Met Lab scientists, were worried about this number. Why not build in a safety factor by adding more tubes, they argued. Other scientists at the Met lab objected. Drilling the extra holes and fitting them with process tubes was a typical example of overconservative engineering, they said. It would slow down the project and delay the delivery of a weapon. Besides, the extra tubes weren’t needed: calculations done at the Met Lab clearly showed that 1,500 tubes were enough. But on this, as on other issues, the engineers in Wilmington had the final say. By the time the blueprints made their way to Hanford, the design contained 2,004 holes for process tubes arrayed in a rectangular pattern.

  Water towers on either side of the B Reactor provided emergency cooling water. Adjacent buildings provided water from the Columbia River and otherwise supported reactor operations. Courtesy of the US Department of Energy.

  Throughout the entire process of designing and building the reactor, security was extremely tight. Construction workers were cleared for one side of the reactor but not the other. They had to produce documentation halfway up a stairway, and only if their names were on a list could they pass. Security was strict for everyone. Enrico Fermi visited Hanford several times during the construction and operation of the reactors. By this time his name was Henry Farmer—all the top scientists in the Manhattan Project had aliases to disguise their identities. Once, when Fermi and Wigner, who had the alias Wagner, were coming through the gate at Hanford, a guard heard Fermi’s accent and asked him if Farmer was his real name. Wigner replied, in his heavy Hungarian accent, “Farmer is his real name just as much as Wagner is my real name.”

  Whenever Fermi came to Hanford, he and Woods tried to find the time to go for a hike and talk. Once, sack lunches in hand, they passed through Hanford’s fence and began exploring the desert. Eventually they settled down beneath an old growth sagebrush to have lunch, discussing how long ago the Columbia River had deposited the sand and gravel on which they sat. They noticed a small plane circling overhead. Soon they saw a man in the distance, holding a revolver but staring at the ground and walking a seemingly random path. He came closer until, as he rounded a nearby sagebrush, he pointed the gun at them and told them to put their hands up. Fermi was holding a sandwich in his hand while Woods held an apple in hers. “With sandwich and apple held high, not daring to take another bite,” she recalled, they were marched at gunpoint back inside the fence.

  EVEN AS THE B REACTOR was nearing completion, a crisis was looming elsewhere at Hanford. The uranium “slugs,” as the workers called the fuel elements, could not be put into the reactor as bare metal. If they were, the water rushing down the process tubes to cool the fissioning fuel would quickly erode the uranium and carry highly radioactive fission products into the Columbia River. The slugs first had to be coated by some kind of metal. But this process of “canning” the fuel elements had become a disaster. DuPont contracted with metallurgists all over the country to come up with ways to do it, but no one could figure out how, and without fuel to load into the reactors the project would fail.

  By this time, DuPont had another source of expertise on which it could draw. To try to anticipate some of the problems it would encounter at Hanford, DuPont had constructed a test reactor at the Clinton Engineer Works in Tennessee, where the Manhattan Project was building the massive plants needed to separate uranium-235 from natural uranium. Larger than Fermi’s Chicago reactor but much smaller than the Hanford reactors, the X-10 Reactor rose on a small hill in the Bethel Valley about 10 miles southwest of the town of Oak Ridge. The X-10 was cooled much more simply than the Hanford reactors—with air pumped through the reactor’s graphite blocks—which made it less useful as a test reactor for Hanford. But the X-10, which went critical on November 4, 1943, provided some early experi
ence with canning fuel elements.

  Still, when Fermi and Woods walked through the slug production line at Hanford on August 1, 1944—with the scheduled startup of the B Reactor less than two months away—they were shocked at how few fuel elements were ready. The problem was getting the metal casing to adhere to the uranium so that it was watertight. The casings tended to leak, rendering the fuel elements worthless. Using the most advanced technologies available, workers in the fuel fabrication plant were producing thousands of fuel elements—and almost all of them were failing the tests they had to pass to be loaded into the reactor.

  This was a problem impervious to all the expertise of the scientists and engineers working on the Manhattan Project. They might be able to calculate how metals should behave in an ideal world, but they had little to no experience with how they actually do behave. The only way to solve the problem was through trial and error. Improvements, even at the last moment, were incremental. Two men figured out how to can and cap the fuel elements in a large pot of molten aluminum—a process that became known as underwater canning. The temperature of the metal baths into which the slugs were dipped had to be set unexpectedly low to get the metal to adhere to the uranium. Workers came up with new methods to test the finished slugs, with the rejects getting melted down and sent back into production.

  The fuel fabrication problem was one of the most visible crises at Hanford, but many such problems, large and small, had to be solved to get the reactors to work. Scientists and engineers provided many solutions, but many came from the machinists, pipefitters, millwrights, carpenters, plumbers, and other craftworkers who built and ran Hanford. Scientists and engineers get much of the credit for innovations in industry and society, but many of those innovations actually have humbler origins. Hanford’s craftworkers were deservedly proud of their contributions to its success. “We worked together to solve problems,” said Cecil Gosney, who capped uranium slugs before they were loaded into the reactors. “Supervision encouraged us by listening to our ideas and suggestions. When we went home we felt we had accomplished something.”