Woods was finishing her thesis work during the summer but sometimes helped Anderson scour Chicago for lumber. CP-1—Chicago Pile Number One—Fermi planned to build in the form of a sphere, the most efficient shape to maximize k. Since the pile’s layers of graphite bricks would enlarge concentrically up to its equator, they would need external support, and wood framing was light and easy to shape and assemble. “I was the buyer for a lot of lumber,” Anderson says. “I remember the Sterling Lumber Company, how amazed they were by the orders I gave them, all with double X priority. But they delivered the lumber with no questions asked. There was almost no constraint on money and priority to get what we wanted.”1667
Horseback riding one Saturday afternoon in the Cook County Forest Preserve twenty miles southwest of Chicago, Arthur and Betty Compton found an isolated, scenic site for the pile building, a terminal moraine forested with hawthorne and scrub oak known as the Argonne Forest. The Army’s Nichols negotiated with the county to use the land; Stone & Webster began planning construction.
The Fermis rented a house from a businessman moving to Washington for war work; since they were enemy aliens and not allowed to own a shortwave radio the man had to have his big all-band Capehart temporarily disabled of its long-distance frequencies, though it continued to supply dance music to the party room on the third floor. Fermi was angry to find his mail being opened and complained indignantly until the practice was stopped (or managed more surreptitiously). The Comptons gave a series of parties to welcome newcomers to the Met Lab. “At each of these parties,” Laura Fermi writes, “the English film Next of Kin was shown. It depicted in dark tones the consequences of negligence and carelessness. A briefcase laid down on the floor in a public place is stolen by a spy. English military plans become known to the enemy. Bombardments, destruction of civilian homes, and an unnecessarily high toll of lives on the fighting front are the result. . . . Willingly we accepted the hint and confined our social activities to the group of ‘metallurgists.’ ”1668 Compton, who describes himself as “one of those who must talk over important problems with his wife,” arranged uniquely to have Betty Compton cleared.1669 None of the other wives was supposed to know about her husband’s work. Laura Fermi found out, like many others, only at the end of the war.
In mid-August Fermi’s group could report a probable k for a graphite-uranium oxide pile of “close to 1.04.”1670 They were working on control-rod design and testing the vacuum properties of both metal sheet and balloon cloth. The cloth was Anderson’s idea, a possible alternative to canning the pile to exclude neutron-absorbing air. It proved serviceable and Anderson followed up: “For the balloon cloth enclosure I went to the Goodyear Rubber Company in Akron, Ohio. The company had a good deal of experience in building blimps and rubber rafts but a square balloon 25’ on a side seemed a bit odd to them.” They made it anyway, “with no questions asked.”1671 It should be good for a 1 percent improvement in k.1672
Between September 15 and November 15 Anderson, Walter Zinn and their crews also built sixteen successive exponential piles in the Stagg Field west stands to measure the purity of the various shipments of graphite, uranium oxide and metal they had begun to receive in quantity. Not all the uranium was acceptable. But Mallinckrodt Chemical Works in St. Louis, specialists at handling the ether necessary for oxide extraction, began producing highly purified brown oxide at the rate of thirty tons a month, and the National Carbon Company and a smaller supplier, by using purified petroleum coke for raw material and doubling furnace time, significantly improved graphite supplies (graphite is molded as coke, then baked in a high-temperature electric-arc oven for long hours until it crystallizes and its impurities vaporize away). By September regular deliveries began to arrive in covered trucks. Physicists doubled as laborers to unload the bricks and cans and pass them into the west stands for finishing.
Walter Zinn took charge of preparing the materials for the pile. The graphite came in from various manufacturers as rough 4¼ by 4¼-inch bars in 17- to 50-inch lengths. So that the bars would fit closely together they had to be smoothed and cut to standard 16½-inch lengths. About a fourth of them also had to be drilled for the lumps of uranium they would hold. A few required slots machined through to make channels for control rods. The uranium oxide needed to be compressed into what the physicists called “pseudospheres”—stubby cylinders with round-shouldered ends—for which purpose the press from the Jersey City junkyard had been shipped to Chicago the previous winter.1673
For crew Zinn had half a dozen young physicists, a thoroughly able carpenter and some thirty high school dropouts earning pocket money until their draft notices came through. They were Back of the Yards boys from the tough neighborhood beyond the Chicago stockyards and Zinn improved the fluency of his swearing keeping them in line.
Machining the graphite was like sharpening thousands of giant pencils. Zinn used power woodworking tools. A jointer first made two sides of each graphite brick perpendicular and smooth; a planer finished the other two surfaces; a swing saw cut the bricks to length. That processing produced 14 tons of bricks a day; each brick weighed 19 pounds.
To drill the blind, round-bottomed 3¼-inch holes for the uranium pseudospheres, two to a brick, Zinn adapted a heavy lathe. He mounted a 3¼-inch spade bit in the headstock of the lathe, where the material to be turned would normally be mounted, and forced the graphite up against the tool with the lathe carriage. Dull bits caused problems. Zinn tried tough carballoy bits first, but they were tedious to resharpen. He began making bits from old steel files, sharpening them by hand whenever they dulled. One sharpening was good for 60 holes, about an hour’s work. Before they were through they would shape and finish 45,000 graphite bricks and drill 19,000 holes.
General Groves made his first appearance at the Met Lab on October 5 and delivered his first pronouncement. The technical council was debating cooling systems again. “The War Department considers the project important,” Seaborg paraphrases Groves’ formula, which they would all learn by heart. “There is no objection to a wrong decision with quick results. If there is a choice between two methods, one of which is good and the other looks promising, then build both.”1674 Get the cooling-system decision into Compton’s hands by Saturday night, Groves demanded. It was Monday. They had been debating for months.
Groves moved on to Berkeley more impressed with their work than his Met Lab auditors realized. “I left Chicago feeling that the plutonium process seemed to offer us the greatest chances for success in producing bomb material,” he recalls. “Every other process . . . depended upon the physical separation of materials having almost infinitesimal differences in their physical properties.” Transmutation by chain reaction was entirely new, but the rest of the plutonium process, chemical separation, “while extremely difficult and completely unprecedented, did not seem to be impossible.”1675
At the beginning of the month, to Compton’s great relief, the brigadier had convinced E. I. du Pont de Nemours, the Delaware chemical and explosives manufacturers, to take over building and running the plutonium production piles under subcontract to Stone & Webster. He meant to involve the industrial chemists more extensively than that—meant for them to take over the plutonium project in its entirety. Du Pont resisted the increasing encroachment. “Its reasons were sound,” writes Groves: “the evident physical operating hazards, the company’s inexperience in the field of nuclear physics, the many doubts about the feasibility of the process, the paucity of proven theory, and the complete lack of essential technical design data.”1676 Du Pont also suspected, once it had sent an eight-man review team to Chicago at the beginning of November, that the plutonium project was the least promising of the several then under development and might even fail, tarnishing the company’s reputation. Nor was it happy at the prospect of identifying itself with a secret weapon of mass destruction; it still remembered the general condemnation it had received for selling munitions to Britain and France before the United States entered the First World War.
Groves told the Du Pont executive committee that the Germans were probably hard at work and the only defense against a Nazi atomic bomb would be an American bomb. And added what he took to be a clinching argument: “If we were successful in time, we would shorten the war and thus save tens of thousands of American casualties.”1677 The second week in November Du Pont admitted the possibility of regular production by 1945 and accepted the assignment (limiting itself to a profit of one dollar to avoid arms-merchant stigma), but made its skepticism and reluctance clear.
By then Stone & Webster’s construction workers had gone on strike. The pile building scheduled for completion by October 20 would be indefinitely delayed. Fermi lived with the problem only long enough to recalculate the risks of pile control. In early November he cornered Compton in his office and proposed an alternative site: the doubles squash court where his team had built its series of exponential piles. A k greater than 1.0 presented an entirely different order of risk from a k of less than 1.0, however; Compton had, in Seaborg’s words, a “dreadful decision” to make.1678 “We did not see how a true nuclear explosion, such as that of an atomic bomb, could possibly occur,” Compton writes with more calm than he probably felt at the time. “But the amount of potentially radioactive material present in the pile would be enormous and anything that would cause excessive ionizing radiation in such a location would be intolerable.”1679 He asked for Fermi’s analysis of the probability of control.
No doubt Fermi discussed the various hand and automatic control rods he planned for the pile. But even slow-neutron fission generations had been calculated to multiply in thousandths of a second, which might flash the pile to dangerous levels of heat and radiation before any merely mechanical control system could move into position. The “most significant fact assuring us that the chain reaction could be controlled,” says Compton, was one of the Richard Roberts team’s earliest discoveries at the Carnegie Institution’s Department of Terrestrial Magnetism following Bohr’s announcement of the discovery of fission in 1939—in Compton’s words, that “a certain small fraction of the neutrons associated with the fission process are not emitted at once but come off a few seconds after fission occurs.”1680 With a pile operating at k only marginally above 1.0, such delayed neutrons would slow the response sufficiently to allow time for adjustment.1681
For once Compton made a quick decision: with control seemingly assured, he allowed Fermi to build CP-1 in the west stands. He chose not to inform the president of the University of Chicago, Robert Maynard Hutchins, reasoning that he should not ask a lawyer to judge a matter of nuclear physics. “The only answer he could have given would have been—no. And this answer would have been wrong. So I assumed the responsibility myself.”1682 The word meltdown had not yet entered the reactor engineer’s vocabulary—Fermi was only then inventing that specialty—but that is what Compton was risking, a small Chernobyl in the midst of a crowded city. Except that Fermi, as he knew, was a formidably competent engineer.
* * *
In mid-November Fermi reorganized his team into two twelve-hour shifts, a day crew under Walter Zinn (who continued to supervise materials production as well), a night crew under Herbert Anderson. Construction began on Monday morning, November 16, 1942. From the balcony of the doubles squash court in the west stands of Stagg Field Fermi directed the hanging of the cubical dark-gray Goodyear balloon as his men hauled it into place with block and tackle. It dominated the room: bottom panel smoothed on the floor, top and three sides secured to the ceiling and the walls, the fourth side facing the balcony furled up out of the way like an awning. Someone drew a circle on the floor panel to locate the first layer of graphite and without ceremony the crew began positioning the dark, slippery bricks. The first layer was “dead” graphite that carried no load of uranium: solid crystalline carbon to diffuse and slow the neutrons that fission would generate. Up the pile as it stacked, the crews would alternate one layer of dead graphite with two layers of bricks each drilled and loaded with two five-pound uranium pseudospheres. That created a cubic cell of neutron-diffusing graphite around every lump of uranium.
To build the wooden framing, Herbert Anderson recalls, “Gus Knuth, the millwright, would be called in.1683 We would show him . . . what we wanted, he would take a few measurements, and soon the timbers would be in place. There were no detailed plans or blueprints for the frame or the pile.” Since they had batches of graphite, oxide and metal of varying purity, they improvised the placement of materials as they went along. Fermi, says Anderson, “spent a good deal of time calculating the most effective location for the various grades of [material] on hand.”1684
They were soon averaging not quite two layers a shift, handing the bricks along from their delivery skids, sliding them to the workers on the pile, singing together to pass the time.1685 The bricks in the dead graphite layers alternated direction, three running east and west and the next three north and south. That gave support to the oxide layers, which all ran together from front to back except at the outer edges, where dead graphite formed an outer shell. The physicist bricklayers had to be careful to line up the slots for the ten control-rod channels that passed at widely distributed points completely through the pile. “A simple design for a control rod was developed,” says Anderson, “which could be made on the spot: cadmium sheet nailed to a flat wood strip. . . . The [thirteen-foot] strips had to be inserted and removed by hand. Except when the reactivity of the pile was being measured, they were kept inside the pile and locked using a simple hasp and padlock, the only keys to which were kept by Zinn and myself.”1686 Cadmium, which has a gargantuan absorption cross section for slow neutrons, held the pile quiescent.
As it grew they assembled wooden scaffolding to stand on and ran loads of bricks up to the working face on a portable materials elevator. Before the arrival of the elevator, during the period when they were building large exponential piles, they had simply leaned over from the precarious 2 by 12-inch scaffolding and reached the bricks up from the men on the floor below. Groves walked in on them one day and dressed them down for risking their necks. The elevator appeared unbidden soon after.
When they achieved the fifteenth layer Zinn and Anderson began measuring neutron intensity at the end of each shift at a fixed point near the center of the pile with the control rods removed. They used a boron trifluoride counter Leona Woods had devised that worked much like a Geiger counter, clicking off the neutron count. Standard indium foils bombarded to radioactivity by pile neutrons gave daily checks on the boron counter’s calibration. Fermi had complained to Segrè in October that he was doing physics by telephone; now he moved a little closer to the work. “Each day we would report on the progress of the construction to Fermi,” Anderson notes, “usually in his office in Eckhart Hall. Then we would present our sketch of the layers that we had assembled and reach some agreement on what would be added during the following shifts.”1687 Fermi took the raw boron-counter and indium measurements and calculated a countdown. As the pile approached its slow-neutron critical mass the neutrons generated within it by spontaneous fission multiplied through more and more generations before they were absorbed. At k= 0.99, for example, each neutron would multiply through an average one hundred generations before its chain of generations died out. Fermi divided the square of the radius of the pile by a measure of the intensity of radioactivity the pile induced in indium and got a number that would decrease to zero as the pile approached criticality. At layer 15 the countdown stood at 390; at layer 19 it dropped to 320.1688 It was 270 at layer 25 and down to 149 at layer 36.
As winter locked down, the unheated west stands turned bitterly cold. Graphite dust blackened walls, floors, hallways, lab coats, faces, hands. A black haze dispersed light in the floodlit air. White teeth shone. Every surface was slippery, hands and feet routine casualties of dropped blocks. The men building the pile, lifting tons of materials every shift, stayed warm enough, but the unlucky security guards stationed at doors and entrances froze. Zinn scavenged rakish
makeshift to thaw them out:
We tried charcoal fires in empty oil drums—too much smoke. Then we secured a number of ornamental, imitation log, gas-fired fireplaces. These were hooked up to the gas mains, but they gobbled up the oxygen and replaced it with fumes which burned the eyes. . . .1689 The University of Chicago came to the rescue. Years before, big league football had been banned from the campus; we found in an old locker a supply of raccoon fur coats. Thus, for a time we had the best dressed collegiate-style guards in the business.
Fermi had originally designed his first full-scale pile as a 76-layer sphere. Some 250 tons of better graphite from National Carbon now promised to reduce neutron absorption below previous estimates; more than 6 tons of high-purity uranium metal in the form of 2¼-inch cylinders began arriving from Iowa State College at Ames, where one of the Met Lab’s chemistry group leaders, Frank Spedding, had converted a laboratory to backyard mass production. “Spedding’s eggs,” dropped in place of oxide pseudospheres into drilled graphite blocks that were then stacked in spherical configuration close to the center of the CP-1 lattice, significantly increased the value of k. Adjusting for the improvements, Fermi saw that they would not need to seal the Goodyear balloon and evacuate the air from the pile and could eliminate some 20 layers: his countdown should converge to zero, k = 1.0, between layers 56 and 57. Instead of a sphere the pile would take the form of a doorknob as big as a two-car garage, a flattened rotational ellipsoid 25 feet wide at the equator and 20 feet high from pole to pole:
Anderson’s crew assembled this final configuration on the night of December 1:
That night the construction proceeded as usual, with all cadmium covered wood in place. When the 57th layer was completed, I called a halt to the work, in accordance with the agreement we had reached in the meeting with Fermi that afternoon. All the cadmium rods but one were removed and the neutron count taken following the standard procedure which had been followed on the previous days. It was clear from the count that once the only remaining cadmium rod was removed, the pile would go critical. I resisted great temptation to pull the final cadmium strip and be the first to make a pile chain react. However, Fermi had foreseen this temptation and extracted a promise from me to make the measurement, record the result, insert all cadmium rods, and lock them all in place.1690