The April 1943 meetings were not the first time the two physicists met. They encountered each other during Fermi’s occasional visits to Berkeley in the 1930s, particularly in 1937 when he spent time learning about Lawrence’s cyclotron and in 1940 when Fermi delivered the Hitchcock Lectures there. Now, though, they would come into increasingly regular contact. In spite of their differences, they developed a deep mutual respect. Oppenheimer respected Fermi’s way of thinking, his way of solving problems, his confidence and the way that confidence inspired others, and of course his enormous contributions to physics prior to the war. Fermi respected Oppenheimer less as a physicist than as the person in whom the US government, to which he now owed total allegiance, had vested authority over the entire Manhattan Project. If he ever disagreed with Oppenheimer, it would never be in public.
In later years, Oppenheimer was subtly dismissive about Fermi: “Not a philosopher. Passion for clarity. He was simply unable to let things be foggy. Since they always are, this kept him pretty active.” And yet he seems to have been profoundly influenced by, and perhaps even obsessed with, his wartime colleague’s intellect and the scale of his contributions to physics. Leona Libby recalls a dinner party in Pasadena after Fermi’s death at which Oppenheimer suggested a parlor game he called “Who do you want to be on your day off?” One would venture a name and the others would try to analyze what the choice signified, in a sort of amateur psychoanalytic fashion. Astonishingly, Oppenheimer chose Fermi. There was silence in the room. No one knew what to ask him. It is significant that Oppenheimer himself chose the game and then chose Fermi, reflecting an inscrutable, deep-seated need to identify with his Manhattan Project colleague.
ONCE OAK RIDGE WAS UP AND RUNNING, ATTENTION TURNED TO the construction of the first major plutonium production reactor at Hanford. Construction at the vast high-security site began in March 1943.
The project was vast, involving the construction of hundreds of buildings, the recruitment of many thousands of laborers, the creation of a virtual state-within-a-state—many times larger in area than Oak Ridge. The centerpiece was three plutonium production reactors, built alongside the river, and a huge chemical separation plant at some distance from the reactors, where spent fuel rods were transported by truck for disintegration and chemical treatment for plutonium extraction. The first reactor to go live was the “B” reactor, followed by the “D” and “F” reactors.
The work was a triumph of organizational collaboration between Groves’s Army Corps of Engineers and Greenewalt’s DuPont engineers. The decision to go ahead and build these behemoths was made after the success of CP-1, but before CP-2 had been completed and well before the Oak Ridge facility went critical in November 1943. Groves was rolling some very heavy dice here, backed by the complete confidence he earned from the civilian leadership of the project and his own faith in his “crackpots” at the Met Lab.
From Chicago, Anderson, the Marshalls, and Wheeler would eventually relocate to Hanford to contribute to the design and construction of the reactors.
Construction on the B reactor began in October 1943, based on a scaled-up version of the Oak Ridge X-10 reactor. The core was significantly larger than the X-10 core, consisting of a graphite block twenty-eight feet wide, thirty-six feet deep, and thirty-six feet high. The face into which fuel rods were fed had channels for 2,004 fuel rods, compared to the X-10’s 1,260 channels. Cadmium and boron safety rods could be controlled automatically in channels running perpendicular to the fuel rods. Fermi and Wigner calculated that in order to produce the volume of plutonium required at the pace demanded by Groves and the team’s leadership, the reactor would have to run very hot indeed—at some 250 megawatts of power, more than some million times the top power achieved by CP-1. This required special treatment of the graphite to cure it against expansion from the ambient heat of the reactions, as well as a more aggressive cooling system to prevent the aluminum cladding of the uranium fuel rods from melting.
The Columbia River provided plentiful fresh water to be pumped through the core’s channels, in the space between each rod and the wall of its channel, and then returned to the river when deemed safe. The creation of 2,004 channels for fuel was an example of overengineering. Neither Fermi nor Wigner thought that so many channels would be necessary, but Greenewalt’s senior engineer, George Graves, in consultation with Wheeler, decided to pack more capacity into the core in case they needed more power. This, it turned out, was perhaps the most fortunate decision made by the DuPont team.
On the occasion of Fermi’s first visit to Hanford, with Wigner in tow, the two were stopped at the gate to the high-security B reactor site. Wigner momentarily forgot that his code name was Gene Wagner and when the guard asked him to identify himself, Wigner said “Wigner—oh, excuse me please, Wagner!” The guard was immediately suspicious. He began to grill the timid, quiet man with the strong Hungarian accent. “Is your name really Wagner?” Fermi immediately intervened. “If his name is not Wagner, my name is not Farmer.” Wigner continues: “And the guard let us pass. That quick self-assurance was so typical of Fermi.”
Fermi arrived at Hanford a week before the B reactor was scheduled to go critical, in mid-September 1944. Greenewalt and Leona Libby escorted him around the vast facility, including the assembly line for the fuel rods that encased the uranium in aluminum/boron sleeves, where technical problems had delayed the start-up of the reactor. Greenewalt never quite got over the excitement of watching Enrico Fermi bring CP-1 to criticality, and Fermi respected the man whose team of engineers could make a large-scale plutonium production plant—reactor, separation facilities, and all the facilities needed to support these—rise from the desert floor in just over a year. They made an incongruous pair, yet the two of them worked well together—so well, in fact, that their partnership lasted long after the war was over.
THE PROBLEMS RELATED TO THE CANNING OF THE URANIUM RODS were finally solved, and by September 26, 1943, the fuel rods were being loaded into the face of the hulking core. Fermi helped in the operation, perched on a scaffold high above the floor, gently easing a rod or two into its channel. The reactor went critical that afternoon and everyone went home to relax as the pile started to generate plutonium. According to Wheeler, the plan was to power the pile up to nine megawatts and then maintain that power level for a while. Then they would run it hotter and hotter until it reached the 250-megawatt level for which it had been designed. Sometime the next morning, however, the reactor began inexplicably to lose power. The operators were puzzled and decided to pull the safety rods further out of the pile to keep the power at a steady nine megawatts, but the power continued to drop. By midafternoon the control rods were almost all the way out of the pile and still the operators were having trouble maintaining nine megawatts. At about four o’clock, Fermi suggested bringing the power down to 400 kilowatts to see if they could hold the reactor there. They couldn’t. By the end of the day, the B reactor—over a thousand tons of graphite and uranium rising thirty-six feet off the ground—was, for all intents and purposes, dead.
FIGURE 18.2. The author at the control panel of B reactor at Hanford. Photo by Susan Schwartz.
Panic now gripped everyone in the reactor building. Over $7 million ($95 million in today’s dollars) had been spent on the reactor alone, not including the staggering cost of the fuel rods. The US government was in the midst of spending what would be $350 million in total, including two other production reactors, separation facilities for the fuel rods, and all the associated facilities for a small city of forty-five thousand people—nearly $5 billion in today’s terms, an unprecedented expenditure by the US government at that time. If this reactor failed to do its job, though, it would not simply be money and time wasted, it would be a crucial logjam in the timeline for production of one of the main materials for the bomb itself. There was enormous pressure to figure out what was going wrong and to do it fast.
Wheeler had spent time over the past year worrying about reactor “poisoning.”
By his calculations, the fission by-products of the controlled chain reaction, particularly at high power, might prove troublesome if allowed to build up. Those by-products could absorb neutrons out of the chain reaction, slowing or even stopping the reaction from working—”poisoning” the reaction. This is why he and DuPont engineer George Graves had decided to “overengineer” the reactor by installing one-third more fuel rod channels than Wigner thought were required for the reactor to work at 250 megawatts.
In any case, it was not immediately clear just what was causing the loss of power. Fermi and Leona Libby suspected that water had leaked into the fuel rods from the cooling system, but upon inspection the water cooling system was intact. There was also concern about a possible leak in pressurized helium being pumped into the graphite core to replace air that might reduce the reproduction factor, but they could find no such leak.
The reactor itself gave the team an important hint at the cause of the problem. Spontaneously, it started up again. Throughout Wednesday afternoon the power returned to the reactor and by Thursday afternoon, September 28, 1943, the reactor reached nine megawatts again, only to falter once more a few hours later.
Wheeler suspected that some radioactive element had been produced as a by-product of the chain reaction, with an enormous ability to absorb neutrons and a fairly brief half-life, on the order of around eleven hours, after which the chain reaction could revive again at full force. He also suspected that because the drop-off occurred only after the reactor was able to reach the fairly high running power of nine megawatts, the real poison was a product of another radioactive by-product that itself did not absorb neutrons but decayed into something that did. Otherwise, the reactor would not have been able to reach nine megawatts in the first place. Wheeler checked a wall chart that listed isotopes that might be created in fission reactions, along with half-lives, looking for a likely suspect. The only one that really seemed to fit the profile was xenon-135, produced in the decay of iodine-135, a known first-generation by-product of uranium fission. Xenon-135 has a half-life of just over nine hours.
He made some rough calculations of the ability of xenon-135 to absorb neutrons. He found that the culprit was produced in about 6 percent of all fission events. He also discovered, to his and everyone else’s astonishment, that its ability to absorb neutrons in its general vicinity was vastly higher than any element previously studied. When Fermi and his Columbia team studied a list of which elements were particularly good neutron absorbers, they had discovered that cadmium was one of the most potent. They had not tested xenon-135, because the isotope was extremely rare and quite unstable. But Wheeler’s calculation suggested that this form of xenon was one hundred thousand times more potent than cadmium.
Fortunately, with xenon’s half-life of only nine hours, the solution was clear. If the reactor was fully loaded with all 2,004 of the uranium fuel channels filled, the reactivity would swamp the effect of the xenon, and the reactor would be able to operate smoothly at its rated power level.
Wheeler and Graves deserve enormous credit for deciding to overengineer reactor B in the event that fission by-product poisoning required an increase in the power of the reactor. In so doing, they allowed the Manhattan Project to keep to its tight deadlines and salvaged the multi-billion-dollar engineering project on which so much depended. When Groves was apprised of the problem, he was furious and lashed out at Compton, whose somewhat feeble response was that the problem would be studied in greater depth at Argonne, where a new pile, CP-3, had been built to supplement CP-2. Groves did not reproach Fermi, though well he might have.
In Fermi’s brilliant career, he demonstrated his fallibility on only a few occasions, none more dramatically perhaps than this one. His 1934 failure to realize that he had split the uranium atom led later to some embarrassment, but we can take some comfort that his failure deprived the fascist government of a four-year head start on nuclear weapons. The reactor B mistake was far more serious and might have led to the loss of the plutonium side of the project altogether. He never commented on his failure to anticipate xenon poisoning, but he must have been enormously embarrassed. Anticipating it would have saved much time, because the adjustments required to swamp the xenon poisoning delayed the running of the reactor at full power for some five months. The B reactor only achieved full power in February 1945, at which time two other reactors, D and F, were nearing completion. Fermi could have forecast this problem, and also could have Wheeler, who had been concerned about poisoning for some time prior to the completion of the reaction and had also not anticipated the xenon problem. At that moment, though, Fermi had other pressing things on his mind. During the summer of 1944, the work at Los Alamos was running into potentially catastrophic roadblocks. Oppenheimer decided that a total reorganization was required and Fermi wound up with another new role, one that would bring him into residence on the secret mesa northwest of Santa Fe.
CHAPTER NINETEEN
ON A MESA
LAURA AND THE CHILDREN WERE THE FIRST FERMIS TO ARRIVE ON the dusty, high-security mesa that was quickly becoming the focus of the Manhattan Project. Her husband was still shuttling between Chicago, Oak Ridge, and Hanford and would not arrive until early September 1944. In midsummer 1944, she took the train, as instructed, from Chicago to Lamy, New Mexico, a town fifteen miles south of Santa Fe. Like all those destined for Los Alamos, she was ignorant of her final destination. Arriving at Lamy, she almost missed her ride into town. A young Army officer was eagerly looking for “Mrs. Farmer.” Laura was not aware of her husband’s code name and at first did not respond. After a few moments of thought, though, she asked if he was looking for “Mrs. Fermi.” The young man looked her over, figured she was indeed the person he was looking for, and brought Laura and the children to 109 East Palace Road in Santa Fe, where the cheerful Dorothy McKibbin dutifully checked her in and gave her and the children ID cards.
From Santa Fe, she was driven along the twenty miles of dirt road, narrow with frequent switchbacks, to the top of the mesa. She settled herself and the children into a modest apartment, on the second floor of a barracks built to Army specifications for standard issue housing. The Fermis could have insisted on more spacious housing along “Bathtub Row,” a street with private houses that was built for project VIPs, where Oppenheimer and his wife Kitty lived, next to Berkeley physicist Edwin McMillan and his wife, Elsie. Those houses were equipped with bathtubs rather than the showers that prevailed in more standard accommodations, hence the name of the street. In typically modest fashion, the Fermis decided not to pull rank and lived where the Army assigned them. Their downstairs neighbors were German theorist Rudolf Peierls and his wife Eugenia. Their old friends the Segrès were also nearby, having moved from Berkeley not long before.
This was the third major upheaval in five years for the Fermi family: first to New York, then to Chicago, and now to Los Alamos. Of the three, the move to Los Alamos was the most dramatic and the most disorienting for the upper-class woman from Rome and her children. They were enclosed in a compound where the highest security procedures prevailed, where the simple act of going into town (Santa Fe, in this case) to buy provisions was difficult and sometimes impossible, and where they knew almost nothing of what was going on around them. The children were not allowed to wander off-site and attended a small school on the grounds. There, Nella and Giulio joined other children of scientists and engineers in elementary school studies. Nella recalls these days as a great adventure; she and the Peierls’ daughter Gaby would sneak out of the compound and then check back in at the main gate, causing considerable consternation. For Nella, it was all very exciting to be set down in the middle of the New Mexican wilderness, with all sorts of important things going on, none of which she understood.
Certain aspects made life at Los Alamos bearable for Laura and the children. Familiar furnishings from the Leonia house, left behind on the assumption that the Fermis’ move to Chicago would be temporary, now arrived to fill the apartment and make it
feel more like home. Many of the families on the mesa were of European origin and were old friends of the Fermis. It must have been pleasing to see the Bethes, and meeting up with the Segrès certainly reminded her of happier days in Rome. There were many others as well. Thrust together in the most extraordinary circumstances, they bonded and supported each other. Laura eventually found work helping the doctor in the Tech Area, the most sensitive section of the facility, where work on the bomb itself was being conducted. She was one of the first people to learn firsthand about the dangers of radiation poisoning. She socialized actively with other wives. Being one of the older women in the group, she was a bit of a mother hen to younger wives whose husbands—straight out of undergraduate school, in many cases—were drafted into the project.
The call for the Fermis to relocate to Los Alamos was perhaps inevitable, but it came specifically in response to a series of major crises in the project. In response to these crises, Groves and Oppenheimer decided to reorganize the Los Alamos project, and Fermi played a key role in that reorganization.
THE MAIN PROBLEM WAS A MATTER OF PHYSICS. THE WORK AT Los Alamos had always assumed that the “gun” method of assembly—shooting one subcritical chunk of metal, either uranium or plutonium, into another subcritical chunk at high velocity so that they would together form a critical mass—would be the most reliable way of creating a fission explosion. All the knowledge developed about U-235 suggested that the gun method would work for the uranium isotope. The initial studies of plutonium, conducted by Seaborg, Segrè, and others at Berkeley with material created in Lawrence’s cyclotron, suggested the same thing. Early on, physicists knew that “implosion”—that is, compressing a subcritical mass of either uranium or plutonium—would also achieve criticality and a fission explosion, but the challenge of actually executing an implosion was daunting.
The Last Man Who Knew Everything Page 27