Разделение изотопов и применение их в ядерном реакторе

Argonne

Shift change at the Y-12 uranium enrichment facility at Oak Ridge.

An Army-OSRD meeting on 25 June 1942 decided to build a pilot plutonium plant in the Argonne Forest southwest of Chicago. In July, Nichols arranged for a lease of 1,000 acres (400 ha) from Cook County, Illinois, and Captain James F. Grafton was appointed Chicago area engineer. It soon became apparent that the scale of operations was too great for the Argonne, and plans were redrawn to locate the pilot plant at Oak Ridge.

Delays in establishing Argonne led Compton to authorize construction of the first nuclear reactor beneath the bleachers of Stagg Field at the University of Chicago. On 2 December 1942 a team led by Enrico Fermi initiated the first artificial[75] self-sustaining nuclear chain reaction in an experimental reactor known as Chicago Pile-1. The point at which a reaction becomes self-sustaining became known as "going critical". Compton reported the success to Conant in Washington D.C. by a coded phone call, saying, "The Italian navigator [Fermi] has just landed in the new world."[76] In January 1943, Grafton's successor, Major Arthur V. Peterson, ordered Chicago Pile-1 dismantled and reassembled at Argonne, as he regarded the operation of a reactor as too hazardous for a densely populated area.

Hanford

By December 1942 there were concerns that even Oak Ridge was too close to Knoxville, Tennessee, in the unlikely but possible event of a major nuclear accident. Groves recruited DuPont in November 1942 to be the prime contractor for the construction of the plutonium production complex. DuPont was offered a standard cost-plus fixed fee contract, but the President of the company, Walter S. Carpenter, Jr., wanted no fee or profit of any kind, and asked for the proposed contract to be amended to explicitly exclude the company from acquiring any patent rights. This was accepted, but for legal reasons a nominal profit of one dollar was agreed upon. After the war, DuPont asked to be released from the contract early, and had to return 33 cents.

Reactor B under construction at the Hanford site

DuPont recommended the site be located far from the existing uranium production facility at Oak Ridge.[79] In December 1942, Groves dispatched Colonel Franklin Matthias and DuPont engineers to scout potential sites. Matthias reported that Hanford Site near Richland, Washington, was "ideal in virtually all respects". It was isolated and near the Columbia River, which could supply sufficient water to cool the reactors which would produce the plutonium. Groves visited the site in January and established the Hanford Engineer Works (HEW), codenamed "Site W". Under Secretary Patterson gave his approval on 9 February, allocating $5 million for the acquisition of 40,000 acres (16,000 ha) of land in the area.[80] The federal government relocated some 1,500 residents of White Bluffs and Hanford, and nearby settlements, as well as the Wanapum and other tribes using the area. A dispute arose with farmers over compensation for crops which had already been planted before the land was acquired. Where schedules allowed, the Army allowed the crops to be harvested, but this was not always possible.[80] The land acquisition process dragged on and was not completed before the end of the Manhattan Project in December 1946

The dispute did not delay work. Although progress on the reactor design at Metallurgical Laboratory and DuPont was not sufficiently advanced to accurately predict the scope of the project, a start was made in April 1943 on facilities for an estimated 25,000 workers, half of whom were expected to live on-site. By July 1944, some 1,200 buildings had been erected and nearly 51,000 people were living in the construction camp. As area engineer, Matthias exercised overall control of the site.[82] At its peak, the construction camp was the third most populous town in Washington state.[83] Hanford operated a fleet of over 900 buses, more than the city of Chicago.[84] Like Los Alamos and Oak Ridge, Richland was a gated community with restricted access, but it looked more like a typical wartime American boomtown: the military profile was lower, and physical security elements like high fences and guard dogs were less evident.[85]

Canadian sites

Cominco had produced electrolytic hydrogen at Trail, British Columbia, since 1930. Urey suggested in 1941 that it could produce heavy water. To the existing $10 million plant consisting of 3,215 cells consuming 75 MW of hydro-electric power, secondary electrolysis cells were added to increase the deuterium concentration in the water from the exchange process from 2.3% to 99.8%. For this process, Hugh Taylor of Princeton developed a platinum-on-carbon catalyst for the first three stages while Urey developed a nickel-chromia one for the fourth stage tower. The final cost was $2.8 million. The Canadian Government did not officially learn of the project until August 1942. Trail heavy water production started in January 1944 and continued until 1956. Heavy water from Trail was used for the Argonne reactor—the first reactor using heavy water and natural uranium—which went critical on 15 May 1944.[86]

The Chalk River, Ontario, site was established to house the Allied effort at McGill University at the Montreal Laboratory. Since the site was 120 miles (190 km) west of Ottawa, a new community was built at Deep River, Ontario, to provide residences and facilities for the team members. The site was chosen for its proximity to the industrial manufacturing of Ontario and Quebec, and access to a rail head adjacent to a large military base, Camp Petawawa. Located on the Ottawa River it had access to abundant water. The first director of the new laboratory was John Cockroft, but he was later replaced by Bennett Lewis. A pilot reactor known as ZEEP (zero-energy experimental pile) became the first Canadian reactor, and the first to be completed outside the United States, when it went critical in September 1945. A larger 10 MW NRX reactor which was designed during the war was completed and went critical in July 1947.[86]

Raw materials

A billboard encouraging secrecy among Oak Ridge workers.

 

Nichols arranged with the State Department for export controls to be placed on uranium oxide and negotiated for the purchase of 1,200 tons of ore from the Belgian Congo in a warehouse on Staten Island. He arranged with Eldorado Mining and Refining for the purchase of ore from its mine in Port Hope, Ontario, and its shipment in 100-ton lots.[87] Mallinckrodt Incorporated in St Louis, Missouri, took the raw ore and dissolved it in nitric acid to produce uranyl nitrate. Ether was then added in a liquid-liquid extraction process to separate the impurities from the uranyl nitrate. This was then heated to form uranium trioxide, which was reduced to highly pure uranium dioxide.[88] By July 1942, Mallinckrodt was producing a ton of highly pure oxide a day, but turning this into uranium metal initially proved more difficult for Westinghouse and Metal Hydrides.[89] Production was too slow and quality was unacceptably low. A special branch of the Metallurgical Laboratory was established at Iowa State College in Ames, Iowa, under Frank Spedding to investigate alternatives. They developed the Ames process, which became available in 1943.

Marshall and Nichols discovered that the electromagnetic process would require 5,000 tons of copper, which was in desperately short supply. However, silver could be substituted, in an 11:10 ratio. On 3 August, Nichols met with Under Secretary of the Treasury Daniel W. Bell and asked for the transfer of 6,000 tons of silver bullion from the West Point Depository. "Young man," Bell told him, "you may think of silver in tons but the Treasury will always think of silver in troy ounces!"[91] Eventually, 14,700 tons would be used. The 1,000-ounce (28,000 g) silver bars were cast into cylindrical billets and taken to Phelps Dodge in Bayway, New Jersey, where they were extruded into strips 0.625 inches (15.9 mm) thick, 3 inches (76 mm) wide and 40 feet (12 m) long. These were wound onto magnetic coils by Allis Chalmers in Milwaukee, Wisconsin. After the war, all the machinery was dismantled and cleaned and the floorboards beneath the machinery were ripped up and burned to recover minute amounts of silver. In the end, only 1/36,000th of a percent was lost.[92][93] The last silver was returned in May 1970.[94]

Uranium

Electromagnetic separation

 

Natural uranium consists of 99.3% uranium-238 and only 0.7% uranium-235, but only the latter is fissile. The rarer but chemically identical uranium-235 has to be physically separated from the more plentiful isotope. Various methods were considered for uranium enrichment, most of which was carried out at Oak Ridge.[95] Electromagnetic isotope separation was developed by Lawrence at the University of California Radiation Laboratory. This method employed devices known as the calutron, which was a hybrid of the familiar laboratory mass spectrometer and cyclotron. The name was derived from the words "California", "university" and "cyclotron".[96] The electromagnetic process was based upon the fact that charged particles are deflected in a magnetic field and the amount of deflection depends upon the particle's mass.[97] The process was neither scientifically elegant nor industrially efficient.[98] It was reckoned that compared with a gaseous diffusion plant or a nuclear reactor, an electromagnetic separation plant would consume more scarce materials, require more manpower to operate and cost more to build. Nonetheless, it was approved because it was based on proven technology and therefore represented less risk. Moreover, it could be built in stages and rapidly reach industrial capacity.[

Giant Alpha I racetrack at Y-12.

Responsibility for the design and construction of the electromagnetic separation plant, which came to be called Y-12, was assigned to Stone & Webster by the S-1 Committee in June 1942. The design called for five first stage processing units, known as Alpha racetracks, and two units for final processing, known as Beta racetracks. In September 1943 Groves authorized construction of four more racetracks, known as Alpha II. Construction began in February 1943.[99] When the plant was started up for testing on schedule in October, the 14-ton vacuum tanks crept out of alignment due to the power of the magnets. As a result, they were fastened more securely. However, a more serious problem arose when the magnetic coils started shorting out. In December Groves ordered a magnet to be broken open, and handfuls of rust were found inside. Groves then ordered the racetracks to be torn down and the magnets sent back to the factory to be cleaned. A pickling plant was established on site to clean the pipes and fittings.[98] As a result, the first Alpha I racetrack was not operational until March 1944. However the second Alpha I was operational by the end of January 1944, the first Beta and third Alpha I came online in March, and the fourth Alpha I was operational in April. The four Alpha II racetracks were completed between July and October 1944.

Operators at their calutron control panels at the Y-12. Gladys Owens, the woman seated in the foreground, did not know what she had been involved with until seeing this photo in a public tour of the facility fifty years later

Tennessee Eastman was hired to manage Y-12 on the usual cost-plus fixed fee basis, with a fee of $22,500 per month plus $7,500 per racetrack for the first seven racetracks and $4,000 per additional racetrack.[101] The calutrons were initially operated by scientists from Berkeley to remove bugs and achieve a reasonable operating rate. They were then turned over to trained Tennessee Eastman operators who had only a high school education. Nichols compared unit production data, and pointed out to Lawrence that the young "hillbilly" girl operators were outperforming his PhDs. They agreed to a production race and Lawrence lost, a morale boost for the Tennessee Eastman workers and supervisors. The girls were "trained like soldiers not to reason why", while "the scientists could not refrain from time-consuming investigation of the cause of even minor fluctuations of the dials."

Y-12 shipped its first few hundred grams of 13 to 15% enriched uranium to Los Alamos in March 1944. Only 1 part in 5,825 of the uranium feed emerged as final product. Much of the rest was splattered over equipment in the process. Strenuous recovery efforts helped raise the enrichment of the product to 10% in January 1945. In February 1945 the Alpha racetracks began receiving slightly enriched (1.4%) feed from the new S-50 thermal diffusion plant. The next month it received enhanced (5%) feed from the K-25 gaseous diffusion plant. By April K-25 was producing uranium sufficiently enriched to feed directly into the Beta tracks.[103]

Gaseous diffusion

The most promising but also the most challenging method of isotope separation was gaseous diffusion. This is based on Graham's law, which states that the rate of effusion of a gas is inversely proportional to the square root of its molecular mass. In a box containing a semi-permeable membrane and a mixture of two gases, the lighter molecules will pass out of the container more rapidly than the heavier molecules. The gas leaving the container is somewhat enriched in the lighter molecules, while the residual gas is somewhat depleted. The idea was that such boxes could be formed into a cascade of pumps and membranes, with each successive stage containing a slightly more enriched mixture. Research into the process was carried out at Columbia University by a group that included Harold Urey, Karl P. Cohen and John R. Dunning.

Oak Ridge K-25 Plant

In November 1942 the Military Policy Committee approved the construction of a 600-stage gaseous diffusion plant.[105] On 14 December, M. W. Kellogg accepted an offer to construct the plant, which was codenamed K-25. A cost plus fixed fee contract was negotiated, eventually totaling $2.5 million. A separate corporate entity called Kellex was created for the project, headed by Percival C. Keith, one of Kellogg's vice presidents. The process faced formidable technical difficulties. The highly corrosive gas uranium hexafluoride would have to be used, as no substitute could be found, and the motors and pumps would have to be vacuum tight and enclosed in inert gas. The biggest problem was the design of the barrier, which would have to be strong, porous and resist to corrosion by uranium hexafluoride. Electro-deposited nickel mesh diffusion barriers were pioneered by Edward Adler and Edward Norris. A six-stage pilot plant was built at Columbia to test the process. Unfortunately, the Norris-Adler prototype proved to be too brittle. A rival barrier was created from nickel powder by Kellex, the Bell Telephone Laboratories and the Bakelite Corporation. In January 1944, Groves ordered the Kellex barrier into production.

Kellex's design for K-25 called for a four-story 0.5-mile (800 m) long U-shaped structure containing 54 contiguous buildings. A survey party began construction by marking out the 500-acre (200 ha) site in May 1943. Work on the main building began in October 1943, and the six-stage pilot plant was ready for operation on 17 April 1944. In 1945 Groves canceled the upper stages of the plant, directing Kellex instead to design and build a 540-stage side feed unit, which became known as K-27. Kellex transferred the last unit to the operating contractor, Union Carbide and Carbon, on 11 September 1945. The total cost, including the K-27 plant completed after the war, came to $480 million.

The production plant commenced operation in February 1945, and as cascade after cascade came online, the quality of the product increased. By April 1945, K-25 had attained a 1.1% enrichment and the output of the S-50 thermal diffusion plant began being used as feed. Some product produced the next month reached nearly 7% enrichment. In August, the last of the 2,892 stages commenced operation. K-25 and K-27 achieved their full potential in the early post-war period, when they eclipsed the other production plants and became the prototypes for a new generation of plants.

Thermal diffusion

Main article: S-50 (Manhattan Project)

The thermal diffusion process was developed by US Navy scientists, and was not one of the enrichment technologies initially selected for use in the Manhattan Project. Then in April 1944, Oppenheimer noted the progress of Philip Abelson's experiments on thermal diffusion at the Naval Research Laboratory and wrote to Groves suggesting that the output of a thermal diffusion plant could be fed into Y-12. Groves set up a committee consisting of Warren K. Lewis, Eger Murphree and Richard Tolman to investigate the idea. They estimated that a thermal diffusion plant costing $3.5 million could enrich 50 kilograms of uranium per week to nearly 0.9% uranium-235. Groves approved its construction on 24 June 1944.

K-25 powerhouse with S-50 plant beyond it, viewed from the air.

Groves contracted with the H. K. Ferguson Company of Cleveland, Ohio, to build the thermal diffusion plant, which was designated S-50. His advisers had estimated that it would take six months to build; Groves gave them just four. Nichols created a special S-50 office within the MED headquarters, headed by Lieutenant Colonel Mark C. Fox, with Major Thomas J. Evans, who would succeed Fox in March 1945, as his deputy. Plans called for the installation of 2,142 48-foot-tall (15 m) diffusion columns arranged in 21 racks. Inside each column were three concentric tubes. Steam obtained from the nearby K-25 powerhouse at a pressure of 100 pounds per square inch (690 kPa) and temperature of 545 °F (285 °C) flowed downward through the innermost 1.25-inch (32 mm) nickel pipe while water at 155 °F (68 °C) flowed upwards through the outermost iron pipe. Isotope separation occurred in uranium hexafluoride between the nickel and copper pipes

 

Work commenced on 9 July 1944 and in September, S-50 began partial operation. Ferguson operated the plant through a subsidiary known as Fercleve. The plant produced just 10.5 pounds (4.8 kg) of 0.852% uranium-235 in October. Leaks limited production and forced shutdowns over the next few months, but in June 1945 it produced 12,730 pounds (5,770 kg).[112] By March 1945, all 21 production racks were operating. Initially the output of S-50 was fed into Y-12, but starting in March 1945 all three enrichment processes were run in series. S-50 became the first stage, enriching from 0.71% to 0.89%. This material was fed into the gaseous diffusion process in the K-25 plant, which produced a product enriched to about 23%. This was, in turn, fed into Y-12, which boosted it to about 85%, sufficient for nuclear weapons.[113]

Centrifuge

The centrifuge process was regarded as the only promising separation method in April 1942. Jesse Beams had developed such a process at the University of Virginia during the 1930s, but had encountered technical difficulties. The process required high rotational speeds, but at certain speeds harmonic vibrations developed that threatened to tear the machinery apart. It was therefore necessary to accelerate quickly through these speeds. In 1941 he began working with uranium hexafluoride, the only known gaseous compound of uranium, and was able to separate uranium-235. At Columbia, Urey had Cohen investigate the process, and he produced a body of mathematical theory making it possible to design a centrifugal separation unit, which Westinghouse undertook to construct.

Scaling this up to a production plant presented a formidable technical challenge. Urey and Cohen estimated that producing a kilogram of uranium-235 per day would require up to 50,000 centrifuges with 1-metre (3 ft 3 in) rotors, or 10,000 centrifuges with 4-metre (13 ft) rotors, assuming that 4-metre rotors could be built. The prospect of keeping so many rotors operating continuously at high speed appeared daunting.[116] Yet when Beams ran his experimental apparatus, he obtained only 60% of the predicted yield, indicating that more centrifuges would be required. Beams, Urey and Cohen then began work on a series of improvements which promised to increase the efficiency of the process. However, frequent failures of motors, shafts and bearings at high speeds delayed work on the pilot plant.In November 1942 the centrifuge process was abandoned by the Military Policy Committee following a recommendation by Conant, Nichols and August C. Klein of Stone & Webster.

Gun-type weapon design

Main article: Little Boy\

A gun-type nuclear bomb.

The Little Boy uranium bomb was a gun-type fission weapon. It worked by mechanically assembling the critical mass from two subcritical masses of uranium-235: a "bullet" and a "target". The chain reaction resulting from collision of the "bullet" with the "target" released tremendous energy, producing an explosion, but also blew apart the critical mass and ended the chain reaction. The configuration of the critical mass determined how much of the fissile material reacted in the interval between assembly and dispersal, and therefore the explosive yield of the bomb. Even a 1% fission of the material would result in a workable bomb, equal to thousands of tons of high explosive. A poor configuration, or slow assembly, would release enough energy to disperse the critical mass quickly, and the yield would be greatly reduced, equivalent to only a few tons of high explosive.

Oppenheimer concentrated the development effort on the gun-type device, which now only had to work with uranium-235 under Lieutenant Commander A. Francis Birch. Birch's group completed the gun-type design, which became Little Boy, in February 1945.[120] The gun-type method was considered so certain to work that no test was carried out before the bomb was dropped over Hiroshima, although an extensive laboratory testing program was undertaken to make sure the fundamental assumptions were correct. The bomb that was dropped used all the existing extremely highly purified uranium-235, and most of the less highly purified material, so there was none available for such a test. The bomb's design was known to be inefficient and prone to accidental discharg

Plutonium

Reactor design

Workers load uranium slugs into the X10 Graphite Reactor

In March 1943, DuPont began construction of a plutonium plant on a 112-acre (45 ha) site at Oak Ridge. Intended as a pilot plant for the larger production facilities at Hanford, it included the air-cooled X-10 Graphite Reactor, a chemical separation plant, and support facilities. Because of the subsequent decision to construct water-cooled reactors at Hanford, only the chemical separation plant operated as a true pilot.[122] The X-10 Graphite Reactor consisted of a huge block of graphite, 24 feet (7.3 m) long on each side, weighing around 1,500 tons, surrounded by 7 feet (2.1 m) of high-density concrete as a radiation shield.[122] The greatest difficulty was encountered with the uranium slugs produced by Mallinckrodt and Metal Hydrides. These somehow had to be coated in aluminum to avoid corrosion and the escape of fission products into the cooling system. The Grasselli Chemical Company attempted to develop a hot dipping process without success. Meanwhile Alcoa tried canning. A new process for flux-less welding was developed, and 97% of the cans passed a standard vacuum test, but high temperature tests indicated a failure rate of more than 50%. The Metallurgical Laboratory eventually developed an improved welding technique with the help of General Electric.

Watched by Fermi and Compton, the X-10 Graphite Reactor went critical on 4 November 1943 with about 30 tons of uranium. A week later the load was increased to 36 tons, raising its power generation to 500 kW, and by the end of the month the first 500 milligrams (7.7 gr) of plutonium was created. Modifications over time raised the power to 4,000 kW in July 1944. X-10 operated as a production plant until January 1945, when it was turned over to research activities.An air-cooled design was chosen for the reactor at Oak Ridge to facilitate rapid construction, but it was recognized that this would be impractical for the much larger production reactors. Initial designs by the Metallurgical Laboratory and DuPont used helium for cooling, before they determined that a water-cooled reactor would be simpler, cheaper and quicker to build. The design did not become available until 4 October 1943; in the meantime, Matthias concentrated on accommodations, improving the roads, building a railway switch line, and upgrading the electricity, water and telephone lines.

Aerial view of Hanford B-Reactor site, June 1944

Work began on the first of three 250 MW reactors, known as Reactor B, on 10 October 1943. Some 390 tons of steel, 17,400 cubic yards (13,300 m3) of concrete, 50,000 concrete blocks and 71,000 concrete bricks were used to construct the 120-foot (37 m) high building. Construction of the reactor itself commenced in February 1944. As at Oak Ridge, the most difficulty was encountered while canning the uranium slugs, which commenced at Hanford in March 1944. They were pickled to remove dirt and impurities, dipped in molten bronze, tin, and aluminum-silicon alloy, and canned using hydraulic presses, and then capped using arc welding under an argon atmosphere. Finally, they were subjected to a series of tests to detect holes or faulty welds. Disappointingly, most canned slugs initially failed the tests, resulting in an output of only a handful of canned slugs per day. But steady progress was made and by June 1944 production increased to the point where it appeared that enough canned slugs would be available to start the reactor on schedule in August.

Watched by Compton, Matthias, Crawford Greenewalt and Fermi, who inserted the first slug, the reactor was powered up beginning on 13 September 1944. Over the next few days, 838 tubes were loaded and the reactor went critical. Shortly after midnight on 27 September, the operators began to withdraw the control rods to initiate production. At first all appeared well but around 03:00 the power level started to drop and by 06:30 the reactor had shut down completely. The cooling water was investigated to see if there was a leak or contamination. The next day the reactor started up again, only to shut down once more. It seemed that the reactor had a half-life of about 9.7 hours.[130][131] Fermi contacted Chien-Shiung Wu, who identified the cause of the problem as neutron poisoning from xenon-135.[132] Fortunately, DuPont engineer George Graves had deviated from the Metallurgical Laboratory's original design in which the reactor had 1,500 tubes arranged in a circle, and had added an additional 504 tubes to fill in the corners. The scientists had originally considered this overengineering and a waste of time and money, but it was found that by loading all 2,004 tubes and carefully monitoring the power levels, the reactor could reach the required power level and efficiently produce plutonium.[133] Reactor D was started on 17 December 1944 and Reactor F on 25 February 1945.[134]