The Manhattan Project, An Interactive History Home The Manhattan Project, An Interactive History Home Department of Energy Home Office of History and Heritage Resources Home DOEHome
J.R. Oppenheimer and General Groves

Schematic of the X-10 Graphite Reactor, Oak RidgePRODUCTION REACTOR (PILE) DESIGN
(Met Lab, 1942)
Events > The Plutonium Path to the Bomb, 1942-1944

By 1942, scientists had established that some of the uranium exposed to radioactivity in a reactor (pile) would eventually decay into plutonium, which could then be separated by chemical means from the uranium.  Important theoretical research on this was ongoing, but the work was scattered at various universities from coast to coast.  In early 1942, Arthur Compton arranged for all pile research to be moved to the Met Lab at the University of Chicago.

Two distinct but related tasks faced the scientists of the Met Lab in early 1942: the construction of a small experimental pile to explore the physics of the fission chain reaction (hence the term "reactor") and planning for a much larger reactor that could produce plutonium on an industrial scale.  Although the two were in theory similar, in practice the much larger production pile would require elaborate controls, radiation shielding, and a cooling system.  In addition, the theory behind the generation and control of chain reactions was still poorly understood, which was why Enrico Fermi's small experimental pile was necessary in the first place.

Drawing of CP-1 Planning for the experimental reactor -- dubbed "CP-1" for "Chicago Pile No. 1" -- began even before Fermi's team from Columbia University arrived in Chicago.  One of the main goals in building CP-1 was to determine the precise value of the neutron reproduction factor "k" for a theoretical reactor of infinite size.  Early experiments leading to a chain-reacting pile were conducted on a squash racquet court under the abandoned west stands of the University of Chicago's football stadium, Stagg Field.  Arthur Compton made plans to build the first pile at a site in the Argonne Forest Preserve, about twenty-five miles southwest of Chicago, "where the hazards would be minimized."  Labor and other difficulties, however, delayed construction at the Argonne site.  Convinced by Fermi that calculations were reliable enough to preclude a catastrophic run-away chain reaction, Compton authorized construction of the pile at the Stagg Field site. Fearing rejection, Compton sought approval for this decision from neither General Groves nor the University of Chicago administration.  Fermi was confident that the world's first nuclear chain reaction could be produced before the end of 1942.

But planning for industrial-scale production piles had to proceed even before the theoretical questions could be explored by CP-1.  The Fermi pile, important as it was, would provide little technical guidance when it came to the complicated cooling, control, and shielding systems required of a large reactor.  Just as would happen with uranium enrichment at Oak Ridge, the job was to design equipment for a technology that was not yet well understood even in the laboratory.

Schematic of a plutonium production reactor at Hanford. In June 1942, a group headed by Arthur Compton's chief engineer, Thomas V. Moore, began designing the first production reactor (pile).  It quickly became clear that a production pile would differ significantly in design from Enrico Fermi's planned experimental reactor (CP-1).  Radiation and containment shielding would be necessary, as would a cooling system.  Although experimental piles like Fermi's did not generate enough power to need cooling systems, any reactor large enough to produce non-trivial amounts of plutonium would have to operate at high power levels and require coolants of some kind.  The Met Lab group considered the full range of gases and liquids as potential coolants in a search to isolate the substances with the best nuclear characteristics, with hydrogen and helium standing out among the gases and water -- even with its marginal nuclear properties and tendency to corrode uranium -- as the best liquid.  In addition, a method was needed for removing the irradiated uranium, preferably without destroying the reactor.  One obvious option was to extend uranium rods into and through the graphite next to cooling tubes. 

During the summer, Moore and his group began planning a helium-cooled pilot pile to be built by Stone & Webster in the Argonne Forest Preserve near Chicago. On September 25, they reported to Compton.  The proposal was for a 460-ton cube of graphite to be pierced by 376 vertical columns containing twenty-two cartridges of uranium and graphite.  Cooling would be provided by circulating helium from top to bottom through the pile.  A wall of graphite surrounding the reactor would provide radiation containment, while a series of spherical segments that gave the design the nickname "Mae West" would make up the outer shell.

Arthur Compton and Vannevar Bush, 1940By the time Compton (left) received Moore's report, he had two other pile designs to consider.  One was a water-cooled model developed by Eugene Wigner and Gale Young, a former colleague of Compton.  Wigner and Young proposed a twelve-foot by twenty-five-foot cylinder of graphite with pipes of uranium extending from a water tank above, through the cylinder, and into a second water tank underneath.  Coolant would circulate continuously through the system, and corrosion would be minimized by coating interior surfaces or lining the uranium pipes.

A second alternative to Mae West was more daring.  Leo Szilard thought that liquid metal would be such an efficient coolant that, in combination with an electromagnetic pump having no moving parts (adapted from a design he and Albert Einstein had created), it would be possible to achieve high power levels in a considerably smaller pile.  Szilard had trouble obtaining supplies for his experiment, primarily because bismuth, the metal he preferred as the coolant, was rare.

General Leslie GrovesOctober 1942 found Leslie Groves (right) in Chicago ready to force a showdown on pile design.  Szilard had been complaining that decisions had to be made so that design could move to procurement and construction.  Compton's delay reflected uncertainty regarding the superiority of the helium pile and awareness that engineering studies could not be definitive until uncertainties surrounding the neutron reproduction factor k had been cleared up, which would not happen until experiments with CP-1 began.  Some scientists at the Met Lab urged that a full production pile be built immediately, while others advocated a multi-step process, perhaps beginning with an externally cooled reactor proposed by Fermi.  The situation was tailor-made for a man with Groves's temperament.  On October 5,  Groves gave the Met Lab one week to decide.  Even wrong decisions were better than no decisions, Groves claimed, and since time was more valuable than money, more than one approach should be pursued if no single design stood out.  While Groves did not mandate a specific decision, his imposed deadline forced the Met Lab scientists to reach a consensus.

Compton decided on compromise. Fermi would explore the precise value of k and study the fundamentals of pile operation in CP-1, to be completed and in operation by the end of the year.  An intermediate pile with external cooling would be built at Argonne and operated until June 1, 1943, when it would be taken down for the extraction of its resulting plutonium. The 100,000-kilowatt helium-cooled Mae West, designed to produce 100 grams of plutonium a day, would be built at Oak Ridge and operating by March 1944.  Compton and the pile researchers hoped that this pile would function as both a test facility and the first unit of the full-scale production plant. Studies on liquid-cooled reactors, meanwhile, would continue, including Szilard's work on liquid metals. Once again, in the absence of one clearly preferable approach, the urgency of the Manhattan Project required that every possibility be explored simultaneously.

   Next


Sources and notes for this page.

The text for this page was adapted from, and portions were taken directly from the Office of History and Heritage Resources publications: F. G. Gosling, The Manhattan Project: Making the Atomic Bomb (DOE/MA-0001; Washington: History Division, Department of Energy, January 1999), 26-27, and Richard G. Hewlett and Oscar E. Anderson, Jr., The New World, 1939-1946: Volume I, A History of the United States Atomic Energy Commission (Washington: U.S. Atomic Energy Commission, 1972), 108-9, 174-82.  Also used were Jack M. Holl, Argonne National Laboratory, 1946-96 (Urbana, IL: University of Illinois Press), 13-16, and Vincent C. Jones, Manhattan: The Army and the Atomic Bomb, United States Army in World War II (Washington: Center of Military History, United States Army, 1988), 190-91. The terms "atomic pile" and "nuclear reactor" refer to the same thing.  The term "pile" was more common during early atomic research, and it was gradually replaced by "reactor" in the later years of the Manhattan Project and afterwards.  In this web site, the phrase "pile (reactor)" is used to refer to early, experimental piles, and "reactor (pile)" is used to refer to later production reactors, which had more elaborate controls and in general more closely resembled post-war reactors.  Much as the term "pile" gradually gave way to "reactor," "atomic" was gradually replaced by "nuclear."  The schematic drawing of X-10 is reproduced from Hewlett and Anderson, The New World, 195.  The drawing of CP-1 is courtesy the National Archives.  The Hanford reactor schematic is reproduced from the Department of Energy report Linking Legacies: Connecting the Cold War Nuclear Weapons Production Processes to their Environmental Consequences (Washington: Center for Environmental Management Information, Department of Energy, January 1997), 164.  The photograph of Vannevar Bush and Arthur Compton is courtesy the Lawrence Berkeley National Laboratory.  The portrait of Leslie Groves is courtesy the Los Alamos National Laboratory.

Home | History Office | OpenNet | DOE | Privacy and Security Notices
About this Site | How to Navigate this Site | Note on Sources | Site Map | Contact Us