|
Site Selection
By the time President Roosevelt authorized the Manhattan Project on December 28, 1942, work on the east Tennessee site where the first production facilities were to be built was already underway.
On Saturday, September 19, Groves had approved the acquisition of 59,000 acres of land along the Clinch River, 20 miles west of Knoxville, Tennessee. Also approved was the removal of relatively few families on the marginal farmland and extensive site preparation to provide the transportation, communications, and utility needs of the town and production plants that would occupy the previously undeveloped area. At first, this location was known as "Site X" and later changed to the Clinton Engineer Works, named after the nearest town. After the war, the name was again changed officially to Oak Ridge.
Original plans called for the military reservation to house approximately 13,000 people in prefabricated housing, trailers, and wood dormitories. By the time the Manhattan Engineer District headquarters were moved from Washington DC to Tennessee in the summer of 1943 (Groves kept the Manhattan project's office in Washington and placed Col. Kenneth D. Nichols in command at Tennessee), estimates for the town of Oak Ridge had been revised upward to 45,000 people. (Note: The name Oak Ridge did not come into usage until after World War II but will be used here to avoid confusion). By the end of the war, Oak Ridge was the fifth largest city in Tennessee and was consuming 1/7 of all the electrical power being produced in the United States. While the Army and its contractors tried desperately to keep up with the rapid influx of workers and their families, services always lagged behind demand.
The three production facility sites were located in valleys away from the town. This provided security and containment in case of accidental explosions. The Y-12 area, home of the electromagnetic plant, was closest to Oak Ridge, being one ridge away to the south. Farther to the south and west lay both the X-10 area, which contained the experimental plutonium pile and separation facilities, and K-25, site of the gaseous diffusion plant and later the S-50 thermal diffusion plant. Y-12 and X-10 were begun slightly earlier in 1943 than was K-25, but all three were well along by the end of the year.
The X-10 Graphite Reactor was a semi-works (pilot plant) facility based on design and engineering information developed at the Metallurgical Laboratory at the University of Chicago. Based on Fermi's pile (CP-1), the X-10 Reactor and associated chemical extraction facility, produced the world's first plutonium outside the laboratory. Technology developed at Oak Ridge formed the foundation upon which the giant plutonium producing facilities at Hanford were based. For further information, check out Met Lab , Hanford and DuPont.
DuPont broke ground for the X-10 complex at Oak Ridge in February 1943. The site would include an air-cooled experimental pile reactor, a pilot chemical separation plant, and various support facilities. Cooper produced blueprints for the chemical separation plants in time for construction to begin in March. A series of huge underground concrete cells, the first of which sat under the pile, extended to one story above ground. Aluminum cans containing uranium slugs would drop into the first cell of the chemical separation facility and dissolve and then begin the extraction process.
The pile building itself went up during the spring and summer of 1943, a huge concrete shell seven feet thick with hundreds of holes for uranium slug placement. Slugs were to plutonium piles what barrier was to gaseous diffusion; that is, an obstacle that could shut down the entire process. ALCOA (Aluminum Company of America) was the only firm left working on a process to enclose uranium 235 within aluminum sheaths, and it was still having problems. Initial production provided mixed results, with many cans failing vacuum tests because of faulty seams.
The moment everyone had been waiting for came in late October 1943 when DuPont completed construction and tests of the X-10 pile. After thousands of uranium slugs were loaded, the pile went critical in the early morning of November 4th and produced its first plutonium by the end of the month. Criticality was achieved with only half of the channels filled with uranium. During the next several months, Compton gradually raised the power level of the pile and increased plutonium yield.
Chemical separation techniques using the bismuth phosphate process were so successful that Los Alamos received its first plutonium samples beginning in the spring of 1944. Fission studies of these samples at Los Alamos during the summer heavily influenced bomb design.
Y-12 Plant
Electromagnetic Separation of Uranium 235
Operator: Tennessee Eastman Company (Kodak)
Although the Lewis Committee report had placed gaseous diffusion ahead of the electromagnetic approach, many were still betting in early 1943 that Lawrence and his mass spectrograph would eventually predominate. Lawrence and his laboratory of mechanics at Berkeley continued to experiment with the giant 184-inch magnet, trying to reach a consensus on which shims, sources, and collectors to incorporate into the Y-12 design for the Oak Ridge plant. Research on magnet size and placement and beam resolution eventually led to a "racetrack" configuration of two magnets with forty-eight gaps containing two vacuum tanks each per building, with ten buildings being necessary to provide the 2,000 sources and collectors needed to separate 100 grams of uranium 235 daily. It was hoped that improvements in calutron design, or placing multiple sources and collectors in each tank, might increase efficiency and reduce the number of tanks and buildings required, but experimental results were inconclusive even as Stone & Webster of Boston, the Y-12 contractor, prepared to break ground.
At a meeting of Groves, Lawrence, and John Lotz of Stone & Webster in Berkeley late in December 1942, Y-12 plans took shape. It was agreed that Stone & Webster would take over design and construction of a 500-tank facility, while Lawrence's laboratory would play a supporting role by supplying experimental data. By the time another summit conference on Y-12 took place in Berkeley on January 13 and 14, Groves had persuaded the Tennessee Eastman Company to sign on as plant operator and arranged for various parts of the electromagnetic equipment to be manufactured by the Westinghouse Electric Company, the Allis-Chalmers Manufacturing Company and the Chapman Valve Manufacturing Company. At the same time, General Electric agreed to provide electrical equipment.
On January 14, after a day of presentations and a demonstration of the experimental tanks in the cyclotron building, Groves stunned the Y-12 contractors by insisting that the first racetrack of ninety-six tanks be in operation by July 1 and that 500 tanks be delivered by year's end. Given that each racetrack assembly was 122 feet long, 77 feet wide and 15 feet high; that the completed plant was to be the size of three, large two-story buildings; that tank design was still in flux; and that chemical extraction facilities also would have to be built, Groves' demands were little less than shocking. Nonetheless, Groves maintained that his schedule could be met.
For the next two months Lawrence, the contractors, and the Army negotiated over the final design. While all involved could see possible improvements, there simply was not enough time to incorporate every suggested modification. Y-12 design was finalized at a March 17 meeting in Boston, with one major modification - the inclusion of a second stage of the electromagnetic process. The purpose of this second stage was to take the enriched uranium 235 derived from several runs of the first stage and use it as the "feed material" for a second stage of racetracks containing tanks approximately half the size of those in the first. Groves approved this arrangement and work began on both the Alpha (first-stage) and Beta (second-stage) tracks.
Groundbreaking for the Alpha plant took place on February 18, 1943. Soon blueprints could not be produced fast enough to keep up with construction as Stone & Webster labored to meet Groves' deadline. The Beta facility was actually begun before formal authorization. While laborers were aggressively recruited, there was always a shortage of workers skilled enough to perform jobs according to the rigid specifications. (A further complication was that some tasks could be performed only by workers with special security clearances). Huge amounts of material had to be obtained (38 million board feet of lumber, for instance), and the magnets needed so much copper for windings that the Army had to borrow close to 15,000 tons of silver bullion from the United States Treasury to fabricate into strips and wind onto coils as a substitute for copper. Treasury silver was also used to manufacture the busbars that ran around the top of the racetracks.
Replacing copper with silver solved the immediate problem of the magnets and busbars, but persistent shortages of electronic tubes, generators, regulators, and other equipment plagued the electromagnetic project and posed the most serious threat to Groves' deadline. Furthermore, last minute design changes continued to frustrate equipment manufacturers. Nonetheless, when Lawrence toured with Y-12 contractors in May of 1943, he was impressed by the scale of operations. Lawrence returned to Berkeley rededicated to the "awful job" of finishing the racetracks on time.
Lawrence and his colleagues continued to look for ways to improve the electromagnetic process. Lawrence found that hot (high positive voltage) electrical sources could replace the single cold (grounded) source in future plants, providing more efficient use of power, reducing insulator failure, and making it possible to use multiple rather than single beams. Meanwhile, receiver design evolved quickly enough in spring and summer 1943 to be incorporated into the Alpha plant. Work at the Radiation Laboratory picked up additional speed in March with the authorization of the Beta process. With Alpha technology far from perfected, Lawrence and his staff now had to participate in planning for an unanticipated stage of the electromagnetic process.
While the scientists in Berkeley studied changes that would be required in the down-sized Beta racetracks, engineering work at Oak Ridge prescribed specific design modifications. For a variety of reasons, including simplicity of maintenance, Tennessee Eastman decided that the Beta plant would consist of a rectangular, rather than oval, arrangement of two tracks of thirty-six tanks each. Factoring this configuration into their calculations, Lawrence and his coworkers bent their efforts to developing chemical processing techniques that would minimize the loss of enriched uranium during Beta production runs. To make certain that Alpha had enough feed material, Lawrence arranged for research on an alternate method at Brown University and expanded efforts at Berkeley. With what was left of his time and money in early 1943 Lawrence built prototypes of both Alpha and Beta units at Berkeley for testing and training operating personnel. Meanwhile Tennessee Eastman, running behind schedule, raced to complete experimental models so that training and test runs could be performed at Oak Ridge.
But in the midst of encouraging progress in construction and research on the electromagnetic process in July came discouraging news from Oppenheimer's isolated laboratory in Los Alamos, set up in 1943 to consolidate work on atomic weapons. Oppenheimer warned that three times more fissionable material would be required for a bomb than earlier estimates had indicated. Even with satisfactory performance of the racetracks, it was now possible that they might not produce enough purified uranium 235 in time.
Lawrence responded to this crisis in characteristic fashion: He immediately lobbied Groves to incorporate multiple sources into the racetracks under construction and to build even more racetracks. Groves decided to build the first four as planned but, after receiving favorable reports from both Stone & Webster and Tennessee Eastman, allowed a four-beam source in the fifth. Convinced that the electromagnetic process would work and sensing that estimates from Los Alamos might be revised downward in the future, Groves let Lawrence talk him into building a new plant - in effect, doubling the size of the Y-12 complex. The new facility, Groves reported to the Military Policy Committee on September 9, would consist of two buildings, each with two rectangular racetracks of ninety-six tanks operating with four-beam sources.
Shakedown at Y-12
During the summer and fall of 1943 the first electromagnetic plant began to take shape. The huge building to house the operating equipment was readied as manufacturers began to deliver everything from electrical switches to motors, valves, and tanks. While construction forces now totaled more than 15,000 people, another 5,000 operating and maintenance personnel were hired and trained. Then between October and mid-December, Y-12 paid the price for being a new technology that had not been put through its paces in a pilot plant. Vacuum tanks in the first Alpha racetrack leaked and shimmied out of line due to unforeseen tremendous magnetic forces, welds failed, electrical circuits malfunctioned, and operators made frequent and costly mistakes. Most seriously, the magnet coils shorted out because of rust and sediment in the cooling oil.
Groves arrived on December 15 and shut the entire racetrack down. The coils were sent back to Allis-Chalmers with hope that they could be cleaned without being dismantled entirely, while measures were taken to prevent recurrence of the shorting problem. The second Alpha track now bore the weight of the electromagnetic effort. In spite of precautions aimed at correcting the electrical and oil related problems that had shut down Alpha 1, the second Alpha fared little better when it started up in mid-January 1944. While all tanks operated at least for short periods, performance was sporadic and maintenance could not keep up with the electrical failures and defective parts. Like its predecessor, Alpha 2 was a maintenance nightmare.
Alpha 2 eventually produced about 200 grams of twelve-percent uranium 235 by the end of February 1944, enough to send samples to Los Alamos as well as feed the new Beta unit but not enough to satisfy estimates of weapon requirements. The first four Alpha tracks did not operate together until April, a full four months behind schedule. While maintenance improved, output was well under previous expectations. The opening of the Beta building on March 11, 1944 led to further disappointment. Beam resolution was so unsatisfactory that a complete redesign was required. To make matters worse, word spread that the K-25 gaseous diffusion process was in deep trouble because of its ongoing barrier crisis. K-25 had been counted upon to provide uranium enriched enough to serve as feed material for the Beta tracks. Now it would be producing such slight enrichment that the Alpha tracks would have to process K-25's material, requiring extensive redesign and retooling of tanks, doors and liners, particularly in units that would be wired to run as hot, rather than as cold electrical sources.
It became clear to Groves that he would have to find a way for a combination of isotope separation processes to produce enough fissionable material for bombs. This meant making changes in the racetracks so that they could process the slightly enriched material that was being produced at K-25. He then concentrated on further expansion of the electromagnetic facilities. Lawrence, seconded by Oppenheimer, believed that four more racetracks should be built to accompany the nine already finished or under construction. Groves agreed with this approach, though he was not sure that the additional racetracks could be built in time to make a difference.
As the K-25 stock continued to drop and plutonium prospects remained uncertain, Lawrence lobbied yet again for further expansion of Y-12, arguing that it provided the only possible avenue to a bomb by 1945. His plan was to convert all tanks to multiple beams and to build, yet again, two more racetracks. By this time even the British had given up on gaseous diffusion and urged acceptance of Lawrence's plan.
Time was quickly running out, and an element of desperation crept into decisions made at a meeting on July 4, 1944. Groves met with the Oak Ridge contractors to consider proposals Lawrence had prepared after assessing once again the resources and abilities of the Berkeley Radiation Laboratory. There was to be no change in the completed racetracks; there simply was not enough time. Some improvements were to be made in the racetracks then under construction. In the most important decision made at the meeting, Lawrence was to throw all he had into a completely new type of "calutron" that would use a thirty-beam source. Technical support would come from both Westinghouse and General Electric, which would cease work on four-beam development. It was a gamble in a high-stakes game, but sticking with the Alpha and Beta racetracks might have been an even greater gamble.
See: 'Enrichment of Uranium"
K-25 Plant
Gaseous Diffusion Separation of Uranium 235
Operator: Union Carbide Company
June 1943
Eleven miles southwest of Oak Ridge on the Clinch River was the site of the planned K-25 Gaseous Diffusion plant upon which so much hope had rested when it was authorized in late 1942. Championed by the British and placed first by the Lewis Committee, gaseous diffusion seemed to be based on sound theory but had not yet produced any samples of enriched uranium 235.
At Oak Ridge, on a relatively flat area of about 5,000 acres, site preparation for the K-25 power plant began in June. Throughout the summer, contractors contended with primitive roads as they shipped in the materials needed to build what became the world's largest steam electric plant. In September work began on the cascade building, plans for which had changed dramatically since the spring. Now there were to be fifty four-story buildings (2,000,000 square feet) in a U-shape measuring a half-mile long by 1,000 feet wide. Innovative foundation techniques were required to avoid setting thousands of concrete piers to support load-bearing walls.
Since it was eleven miles from the headquarters at Oak Ridge, the K-25 site developed into a satellite town. Housing was supplied, as was a full array of service facilities for the population that eventually reached 15,000. Dubbed "Happy Valley" by the inhabitants, the town had housing similar to that in Oak Ridge, but, like headquarters, it too experienced chronic shortages. Even with a contractor camp with facilities for 2,000 workers nearby, half of the construction force had to commute to the site daily.
In late summer of 1943 it was decided that K-25 would play a lesser role than originally intended. Instead of producing fully enriched uranium 235, the new gaseous diffusion plant would provide around fifty-percent enrichment for use as feeder material for Y-12. This would be accomplished by eliminating the more troublesome upper part of the cascade. Even this level of enrichment was not assured since a suitable barrier for the diffusion process still did not exist. The decision to downgrade K-25 was part of the larger decision to double the capacity of Y-12 and fit with Groves' new strategy of utilizing a combination of separation methods to produce enough fissionable material for bombs as soon as possible.
There was no doubt in Groves' mind that gaseous diffusion still had to be pursued vigorously. Not only had major resources already been expended on the program, but there was also the possibility that it might yet prove successful. Y-12 was in trouble as 1944 began, and the plutonium pile projects (X-10) were just getting underway. A workable barrier design might put K-25 ahead in the race for the bomb. Unfortunately, no one had been able to fabricate barrier material of sufficient quality. The only alternative remaining was to increase production enough to compensate for the low percentage of barrier that met specifications. As Lawrence prepared to throw everything he had into a thirty-beam source for Y-12, Groves ordered a crash barrier program, hoping to prevent K-25 from standing idle as the race for the bomb continued.
S-50 Plant
Liquid Thermal Diffusion Separation of Uranium 235
Operator: Union Carbide Company
September 27, 1944
As problems with both Y-12 and K-25 reached crisis proportions in the spring and summer of 1944, the Manhattan Project received help from an unexpected source - the United States Navy. President Roosevelt had instructed that the atomic bomb effort be an Army program and that the Navy be excluded from deliberations. Navy research on atomic power, conducted primarily for submarines, received no direct aid from Groves, who, in fact, was not up-to-date on the state of navy efforts when he received a letter on the subject from Oppenheimer in April 1944.
Oppenheimer informed Groves that Philip Abelson's experiments on thermal diffusion at the Philadelphia Navy Yard deserved a closer look. Abelson was building a plant to produce enriched uranium to be completed by early July 1944. It might be possible, Oppenheimer thought, to help Abelson complete and expand his plant and use its slightly enriched product as feed material for Y-12 until the problems plaguing K-25 could be resolved.
The liquid thermal diffusion process had been evaluated as early as 1940 by the Uranium Committee, when Abelson was still with the National Bureau of Standards. In 1941 he moved his research to the Naval Research Laboratory, where there was more support for his work. During the summer of 1942 Bush and Conant received reports about Abelson's research but concluded that it would take too long for the thermal diffusion process to make a major contribution to the bomb effort, especially since the electromagnetic and pile projects were making satisfactory progress. After a visit with Abelson in January 1943, Bush encouraged the Navy to increase its support of thermal diffusion. A thorough review of Abelson's project early in 1943, however, concluded that thermal diffusion work should be expanded but should not be considered as a replacement for gaseous diffusion, which was better understood theoretically. Abelson continued his work independently of the Manhattan Project. He obtained authorization to build a new plant at the Philadelphia Navy Yard, where construction began in January 1944.
(Webmaster's Note: The Navy, and specifically the Philadelphia Navy Yard, was chosen by Abelson because of their experience dealing with huge ship boilers which produced steam. Steam was the essential source of heat required for the liquid thermal diffusion process.)
Groves immediately saw the value of Oppenheimer's suggestion and sent a group to Philadelphia to visit Abelson's facility. A quick analysis demonstrated that a thermal diffusion plant could indeed be built at Oak Ridge and placed in operation by early 1945. The steam required in the convection columns was already at hand in the form of the almost completed K-25 powerplant (The largest in the world). It would be relatively simple to provide steam to the thermal diffusion plant and produce enriched uranium, while providing electricity for the K-25 plant when it was finished. Groves gave the contractor, the H. K. Ferguson Company of Cleveland, just ninety days from September 27 to bring a 2,142 column plant on line (In comparison, Abelson's plant in Philadelphia contained 100 columns). There was no time to waste as Happy Valley in Oak Ridge braced itself for a new influx of 10,000 workers.
A Booklet for Oak Ridge Residents
This informational booklet, published by the Roanne-Anderson Company in 1945, was meant to serve as a guide for those living in the restricted military area at Oak Ridge. It reminded residents that their work and cooperation were vital to the war effort, and advised them that "a safe rule to follow is that What you do here, What you see here, What you hear here, please let it stay here." The full booklet can be found here.
|
