Chapter 2: Establishing the Manhattan District
Undeterred by the unusual nature of the atomic energy program, the Army Corps of Engineers in June 1942 prepared to carry out its new wartime construction assignment. After his initial conference with Brig. Gen. Wilhelm D. Styer late in the afternoon of the eighteenth, Col. James C. Marshall experienced a certain restlessness as he tried to comprehend the scope of the new task at hand. The next day, he received some of the answers to his many questions when Styer took him to the Office of Scientific Research and Development to call on Vannevar Bush. The OSRD director gave the two officers several documents, among them a copy of the program for continued development of atomic energy that President Roosevelt had approved on the seventeenth. From these papers Marshall learned that the Army was now charged with “all large-scale aspects,”1 as Bush put it, of the atomic energy program, with the OSRD retaining responsibility for scientific research and pilot plant experimentation. The Army’s mission included building both pilot and full-scale plants for producing fissionable materials to be used in the manufacture of atomic bombs, letting contracts for these plants and others to be under OSRD direction, and extensive site selection, acquisition, and development – all to be carried out in close coordination with the OSRD.
That afternoon, again in General Styer’s office, Colonel Marshall received formal orders on the Army’s phase of the atomic energy project. On the covering letter of the approved program, Styer wrote the following endorsement to Marshall: “This is referred to you for information and appropriate action in accordance with our discussion of this subject with Dr. Bush this morning.”2 This simple statement constituted the basic directive to the Corps of Engineers for its work on development of the atomic bomb. Styer also emphasized that the orders had come directly from the War Department’s Services of Supply (SOS) and that Colonel Marshall would furnish all details of the new project to the Engineers chief, Maj. Gen. Eugene Reybold.
In the weeks that followed the hurried orientation of the past two days, Colonel Marshall became more familiar with the current status of the program
Brig. Gen. Wilhelm D. Styer (1941 photograph)
and what the Army’s role was to be in the months ahead. He was to have broad authority to use engineer facilities, choose personnel, and take whatever steps were necessary to carry out his assignment. Marshall soon realized, however, that he was going to need all the assistance he could muster in order to have any hope of success in achieving his mission.3
Organizing the District
The Engineers chief normally oversaw construction projects through an engineer district, the basic unit of the engineer field organization for supervising construction work. The district engineer customarily was responsible to a division engineer, who headed one of the eleven geographical divisions in the United States (which, in 1942, constituted regional administrative headquarters of the Engineer Department of the Corps). Because of the special character, scope, and importance of Colonel Marshall’s mission, however, the new district to oversee atomic energy construction would be directly subordinate to the Engineers chief and, unrestricted by geographical limitations, its field of operations would extend into other districts and divisions. Furthermore, although designated a district engineer, Marshall was to have all the authority, responsibility, and independence regularly granted to a division engineer. Indeed, in many respects, he was to have far more.4
While Marshall’s responsibility was to the Engineers chief, in practice he worked with Reybold’s assistant, Brig. Gen. Thomas M. Robins, who was in charge of construction, and particularly with his deputy, Col. Leslie R. Groves. During the summer of 1942, Robins and Groves reviewed Marshall’s plans and furnished him with the support and assistance necessary to get the project started. Appropriate agencies of the chief’s staff also cooperated fully with Marshall, who was able to make good use of other engineer facilities and War Department assistance. On all important decisions,
Brig. Gen. James C. Marshall (1946 photograph)
Marshall consulted with Generals Reybold and Styer; the latter, in addition to his many duties as SOS chief of staff, kept well abreast of current nuclear developments. To enforce strict secrecy, Army Chief of Staff General George C. Marshall originally had forbidden Styer to reveal to the SOS commander, Lt. Gen. Brehon B. Somervell, anything about the atomic energy program. In June, however, with the entrance of the Army into an active role in the project, General Marshall directed Styer to brief Somervell and to enlist his support.5
In late June, Colonel Marshall opened a liaison office in Washington, D.C., in the New War Department Building at 21st Street and Virginia Avenue, NW. At the same time, he set up temporary district headquarters at 270 Broadway in New York, where he had ready access not only to the colocated administrative facilities of the Engineer Department’s North Atlantic Division but also to the Manhattan office of the Stone and Webster Engineering Corporation, soon to become a major contractor for the atomic project. To staff the district, Colonel Marshall received authorization from the Engineers chief to draw on officers and civilians who had served under him in the Syracuse District, among them Lt. Col. Kenneth D. Nichols, whom he appointed assistant district engineer. The Syracuse District recently had completed the major part of its wartime construction program and, as the volume of work decreased during the summer, Marshall was able to draw more and more personnel from his former command. Soon over a dozen men had transferred to the new district. Several who were civilians at the time subsequently received reserve commissions and went on active duty. To provide still more officers, General Robins directed other districts to give Marshall a priority on any surplus personnel they might have.6
Col. Kenneth D. Nichols (1945 photograph)
Engineer districts normally took their names from the city where they were located, but Colonel Marshall’s new district lacked a permanent headquarters. Some convenient designation was needed, however, that would conceal the real nature of the project. On 26 June, Generals Somervell, Styer, and Reybold agreed on the elaborate cover name of Laboratory for the Development of Substitute Materials, or DSM. Within the next two weeks Marshall’s plans and organization for a new district were approved and he submitted to Colonel Groves the draft of a general order establishing a DSM District. To Groves, the term DSM seemed likely to arouse attention and curiosity. Accordingly, the two officers reached agreement that the name Manhattan, where Marshall had established his temporary headquarters, would be a better name. On 13 August, General Reybold issued a general order (effective on the sixteenth) officially establishing “a new engineer district, without territorial limits, to be known as the Manhattan District, … with headquarters at New York, N.Y., to supervise projects assigned to it by
the Chief of Engineers.”7 The term DSM continued in use as an official code name for the entire project, but the word Manhattan – symbolically representing the Army’s contribution in the development of the atomic bomb – gradually superseded it.8
Meanwhile, Vannevar Bush carried out the necessary changes in the OSRD organization. Under the provisions of the newly approved program, the OSRD retained responsibility for pilot plants for the centrifuge, diffusion, and electromagnetic separation processes, as well as for further research and development on the latter method, for the heavy water project, and for additional miscellaneous research. Acting upon a suggestion of James B. Conant, based upon his review of past operating procedures of the uranium project, Bush abolished the S-1 Section and its planning board and, in their place, established the S-1 Executive Committee. Membership of the new committee included most of the individuals who had previously served in the S-1 Section: Conant, as chairman; Lyman J. Briggs; Ernest O. Lawrence; Arthur H. Compton; Harold C. Urey; and Eger V. Murphree, with the addition of Irvin Stewart, the OSRD executive secretary. H. T. Wensel, formerly of the planning board, became technical aide. Only Dean George B. Pegram of Columbia dropped out.
Bush charged the new S-1 Executive Committee with recommending contracts and supervising contract operations and enjoined its members to begin work immediately, in close coordination with the Army Corps of Engineers. Bush particularly cautioned them on the importance of maintaining the “greatest secrecy” on all phases of the project, and stated that “we will continue ... to adhere to the principle that confidential information will be made available to an individual only insofar as it is necessary for his proper functioning in connection with his assigned duties.”9
An additional, though temporary, responsibility of the S-1 Executive Committee was overseeing experimentation on the military applications of atomic energy. As outlined in the atomic energy program approved by the President, the Joint Committee on New Weapons and Equipment of the Joint Chiefs of Staff had primary responsibility for this administrative mission. Vannevar Bush also headed this committee. Serving with him were Brig. Gen. Raymond G. Moses, chief of the Supply Division (G-4) of the Army General Staff, and Rear Adm. Willis A. Lee, Jr., who held a similar position as Assistant Chief of Staff (Readiness), U.S. Fleet. With these officers Bush raised the question of establishing a subcommittee to consider military uses of atomic energy – formed, not hastily, but with “great care.- Pending organization of this new group, Bush directed the S-1 Executive Committee to continue its work on military applications.10
The relationship between the Manhattan District and the OSRD S-1 Committee during the summer of 1942 can best be described as a cooperative one. While each agency had its assigned functions within the overall atomic energy program, they coordinated either formally or informally on all major decisions. But they did not act together as a joint directorate, for each organization was free to proceed as it wished to carry out decisions, or other activities, strictly within its own area of competence.
Periodic meetings of the S-1 Committee with Colonel Marshall and one or more other officers of the Manhattan District provided the formal link between the two organizations. Representatives of the principal engineering or industrial firms connected with the project also attended frequently. During this period, the S-1 Committee met at least once a month, usually in executive session in the morning – while Marshall was conferring with his military superiors – and then opened the meeting to the Manhattan representatives. These joint meetings encouraged a free exchange of views, provided scientific briefings for Marshall and his colleagues, enabled the scientists to seek Army assistance where necessary, and generally enhanced coordination.11
Army-OSRD links were further strengthened by cooperation between Manhattan officers and civilian scientists working together on specific projects. In the beginning these ties were kept inconspicuous, especially to conceal the Army’s interest. In their visits to university or industrial laboratories, Army officers usually wore civilian clothing, and every effort was made to hide the relationship between the Corps of Engineers and OSRD-directed projects. This effort was sometimes frustrated when a few scientists, unaccustomed to working under rigid security conditions, talked more freely than they should have about the Army’s interest in their work. And despite Bush’s warnings, even the S-1 Committee was careless on occasion. In mid-August, for example, Colonel Marshall had to point out that highly classified material should not be sent to him through the regular mail. In general, however, the good relationship between the Manhattan District and the S-1 Committee helped to keep such occurrences to a minimum.
Details of the Army-OSRD meetings reached the Top Policy Group through twin channels: scientific and military. Conant reported to Bush and Colonel Marshall to his superiors in General Reybold’s office, or sometimes directly to General Styer. The latter then passed on information about the atomic project to Generals Somervell and Marshall. Secretary of War Henry L. Stimson appears to
have received only limited data on developments during the summer of 1942; Vannevar Bush submitted only one formal report to Harvey Bundy, the Secretary’s special assistant for scientific affairs. To what extent Vice President Henry A. Wallace received information on atomic developments is unclear; the Top Policy Group did not meet during this period and there is no other indication that reports were sent to Wallace. Even the President’s information and activities were evidently limited to the question of nuclear collaboration with Great Britain, and he seems to have discussed that only with Bush. In effect, then, the S-1 Executive Committee and Manhattan District were free to act on any mutually approved decision. Their scientific or military superiors could always exercise the right of veto, but in the summer of 1942 they apparently did not do so. Only later, when major changes were to be made in the atomic energy program, would they once more actively enter the picture.12
Army-OSRD Planning Meeting, 25 June 1942
On the occasion of the first meeting of the S-1 Executive Committee, convened at the Carnegie Institution in Washington, D.C., on 25 June 1942, General Styer, Colonels Marshall and Nichols, Vannevar Bush, and the regular members of the committee reached several important decisions regarding site selection, contracting with engineering firms, and obtaining government priorities for needed materials and equipment.13
War Department policy normally required location of new munitions plants out of range of enemy carrier-based planes, in a great inland zone between the Appalachian and Rocky Mountains and approximately 200 miles from the nation’s borders with Canada and Mexico.14 General Styer stated that the main atomic energy installations should be placed within this zone and that, to ensure secrecy, all manufacturing plants should be built at a single site. The group generally agreed with Styer on plant concentration, which would enable rapid and economical construction and facilitate control over the work. To support the extensive facilities, a continuous supply of approximately 150,000 kilowatts of electricity would be needed by the end of 1943 and hundreds of thousands of gallons of water per minute. There would have to be a climate suitable for construction in winter, a ready supply of labor, an accessibility to transportation, a relative immunity from enemy attack, and a terrain cut up by ridges that would limit the effects of any accidental explosion.
Some steps for finding a satisfactory site already had been taken. An OSRD-directed study group in early April had picked out an area near Knoxville, close to the region under intensive development by the Tennessee Valley Authority,
as suitable for the full-scale centrifuge and diffusion separation plants. (See Map 1.) About the same time, members of Arthur Compton’s team at the Metallurgical Laboratory in Chicago had been seeking a site for the full-scale plutonium production plant. They seriously weighed the possibilities of two locations near Chicago, but finally concluded that the Tennessee Valley was also the best area for their purposes. In mid-June, Bush expressed his liking for the Tennessee site to General Styer, and Colonel Marshall, in one of his first moves as district engineer, also discussed its merits with Colonel Groves. Groves made a quick survey of the electric power situation and indicated his approval of the Knoxville area. Thus, Army representatives recommended the Tennessee Valley location for all the large-scale production plants.15
All scientific leaders at the 25 June planning meeting accepted this recommendation save Lawrence, who maintained that the electromagnetic separation plant ought to be located closer to his research operations in California. Bowing to his objections, the conferees agreed to postpone a decision on location of the electromagnetic plant, pending further progress in basic research on this process. Even though research for the centrifuge and diffusion methods was still at a stage where firm planning for production installations was impractical, the group decided that the plants for these processes as well as for the plutonium process would be located on a 200-square-mile site in the Tennessee Valley. The Army, the planning group agreed, should begin steps at once to select and acquire this site.
The planners also considered sites for two other operations. The first was a pilot plutonium plant required by the Metallurgical Laboratory. This plant needed to be within commuting distance of the laboratory; but, for reasons of safety and security, it could not be built in heavily populated Chicago. Consequently, Compton and his colleagues selected an isolated area known as the Argonne Forest, a part of the Cook County Forest Preserve about 20 miles southwest of the city. This selection was tentatively approved on 25 June and the next day Compton and Colonel Nichols reached final agreement on the general plan for the Argonne site.16
In 1941, the OSRD had sponsored laboratory tests at Princeton University, under the direction of British chemist Hugh S. Taylor, to develop a technique for large-scale production of heavy water by a hydrogen-water exchange process. Taylor had found that this process operated most efficiently when using the electrolytic method to produce hydrogen. The Consolidated Mining and Smelting Company, a Canadian firm, operated an ammonia plant at Trail on the Columbia River, situated a few miles
Projected site for Atomic Production Plants, Tennessee, 1942
north of the Canadian-U.S. border, that was the largest producer of hydrogen by the electrolytic method in North America. In an effort to tap this resource for heavy water, Taylor met with company officials to discuss the possibility of Consolidated Mining allowing its plant to be altered so that heavy water could be extracted from the hydrogen supply without using up any appreciable quantities of the hydrogen. The reaction was favorable. The OSRD therefore made the final arrangements and, in May of 1942, engaged the Boston construction firm of E. B. Badger and Sons to proceed with the engineering and design work on plant alterations. During the 25 June meeting, Army-OSRD representatives approved the plan for Trail and, on the twenty-sixth, shifted responsibility for construction to the Army but left the financing and direction of research with the OSRD.17
A few days before the meeting, Colonel Marshall had talked with Groves about his urgent need for competent engineering advice in organizing the atomic project and Groves had recommended Stone and Webster. The well-known Boston firm already was involved in an OSRD project on the diffusion method and was currently maintaining a good record on several contracts for the Corps of Engineers. Marshall proposed to the Army-OSRD group that it engage Stone and Webster as architect-engineer-manager for the atomic project, to monitor site development and major construction.18
The Army-OSRD conferees approved Marshall’s proposal and agreed that Stone and Webster would be primarily concerned with site development and housing construction in ‘Tennessee and with engineering and building the centrifuge and electromagnetic plants. In addition, the firm would build the Argonne pilot plant and, eventually, the full-scale plutonium production plant. The group also decided to engage the M. W. Kellogg Company of New Jersey. This firm had extensive experience in the design and construction of petroleum refineries and chemical installations and was already assisting the scientific team at Columbia on diffusion research under an OSRD contract. Kellogg would take responsibility for the diffusion plant and Badger and Sons would continue on the job at Trail.19
The Army-OSRD group decided that a substantial number of OSRD research contracts already in operation should be extended at the discretion of the S-1 Executive Committee. To continue these contracts beyond the end of the fiscal year – less than a week away – the atomic program urgently needed $15 million. Marshall promised to obtain the money immediately from engineer funds. This sum represented slightly less than half of the $31 million included for the OSRD in the program
approved by President Roosevelt on 17 June. Yet it sufficed, for the Army gradually took over most OSRD functions in the field of atomic energy and the $15 million proved to be more than enough to finance all further OSRD expenditures for the atomic program.20
The final subject at the Army-OSRD meeting was the urgent need to obtain government priorities sufficiently high to ensure a ready supply of critical materials and equipment. Some required items were in extremely short supply and the OSRD was having little success obtaining them. What was needed, OSRD representatives told Colonel Marshall, was some means of coordinating their requirements and gaining the necessary priorities to satisfy them. They requested that the Army designate a priorities officer to meet with them and to establish an office in Washington, D.C. Marshall agreed and indicated that, as a first step, he would eliminate obvious competition by coordinating Army and OSRD procurement on the atomic project.
The decisions of the Army and OSRD representatives had served to inaugurate officially a new phase in the atomic energy program, a period of Army-OSRD cooperation that would last until late spring of 1943.
Progress in Research and Development
For Colonel Marshall and his Manhattan District associates, the summer of 1942 was a period of organization and planning to lay groundwork for developing an unprecedented weapon. The scope of the problem was broad. To the normal administrative headaches of setting up a very large construction and manufacturing project were added the problems of expediting and coordinating research, experimentation, industrial application, and design of a weapon based on materials that in all probability would not be available for testing until the weapon itself had been built. Of all the problems to be dealt with – the execution of engineering, construction, and operating contracts; the selection and acquisition of sites; the obtaining of large sums of money and of adequate priorities; the procurement of materials; the maintenance of security – the first and most basic to the success of the whole project was that of continued progress in scientific development. On this rested the outcome of the entire atomic enterprise. And, in mid-1942, Marshall found that nearly all problems he faced were connected in one way or another with the vital task of research and experimentation.
The program adopted on 17 June called for backing all five methods of producing fissionable materials – until one or more proved most certain of success. Thus, each process was subject to intensive research efforts that summer. Objectives for the centrifuge process were a pilot plant and, by January 1944, a full-scale plant producing 100 grams of U-235-enriched uranium per day. A low-yield pilot plant and a 1-kilogram-per-day production plant were planned for the diffusion process and a 5-gram-per-day pilot plant and, by late 1943, a 100-gram-per-day production plant for the electromagnetic process.
The plutonium project needed a 100- gram-per-day production plant, as well as heavy water plants producing 0.5 tons per month by May 1943.21 The diffusion and centrifuge methods, which had seemed most promising the previous autumn, now appeared less certain than the others. And of the pile and electromagnetic processes, the latter appeared to offer the best immediate hope.
Although the feasibility of the gaseous diffusion method had been demonstrated, two major problems stood in the way of achieving large-scale separation of uranium isotopes. First was development of a material that would be sufficiently porous to permit passage of uranium hexafluoride through thousands of stages, as well as be resistant to the exceptional corrosiveness of this gas and suitable for fabrication by mass production methods. The second was to design and manufacture corrosion-resistant mechanical equipment – a variety of pumps, valves, seals, and instruments – to move the gas through miles of pipe, maintaining required vacuum conditions, temperatures, and pressures.
These problems were being studied mainly at Columbia University under John Dunning and Harold Urey, but also by the M. W. Kellogg Company, whose chief concern was major production of diffusion equipment and, eventually, construction of a full-scale plant. At a meeting of the OSRD S-1 Executive Committee on 30 July 1942, Urey reported his hope that the diffusion method would be producing enough enriched uranium by the fall of 1944 to begin using that material in an atomic weapon.22
Work on the centrifuge process was going equally slow. Under the general direction of Eger Murphree, theoretical and experimental research continued at Columbia University and the University of Virginia, respectively; design and development at the Westinghouse Research Laboratories, a subsidiary of the Westinghouse Electric and Manufacturing Company; and engineering studies at the Standard Oil Development Company. Feasibility of the method had long been demonstrated, but major technical and mechanical difficulties prevented rapid progress. Nevertheless, a pilot plant had been designed at Standard Oil and actual production of parts and models for the pilot plant was under way at Westinghouse. Like the diffusion process, the centrifuge process would require many hundreds of stages to achieve large-scale separation. Also by Murphree’s estimate, the centrifuge method could not produce a sufficient amount of enriched uranium for use in atomic weapons before autumn of 1944.23
Of all the programs in progress that summer, perhaps the most extensive was the pile process for manufacturing plutonium under the leadership of Arthur Compton at the University of Chicago. The objectives of the Chicago program were to prove experimentally that a chain reaction
was actually possible, then to devise a means to produce plutonium on a large scale and extract it chemically from uranium, to work out the necessary data for bringing about an explosive chain reaction with either plutonium or U-235, and, finally, to design the atomic bomb itself. Investigations into all of these problems were being conducted simultaneously by large research staffs at the University of Chicago and other institutions.24 One particularly important group at the University of California (Berkeley) had been organized in June by J. Robert Oppenheimer, then widely regarded as “the leader of theoretical aspects of atomistics and similar subjects of physics.”25 Under Oppenheimer’s direction a number of the nation’s ablest theoretical physicists undertook a study that, in Oppenheimer’s words, “for the first time really came to grips with the physical problems of atomic bombs, atomic explosions to initiate thermonuclear reactions.” By the latter he meant the possibility of a hydrogen bomb, a matter that he raised with Compton and Bush that summer and that was to lie heavy on his mind for many years to come.26
The University of Chicago’s Metallurgical Laboratory staff also devoted most of its energies to theoretical studies. Lack of basic materials – uranium, plutonium, highly purified graphite, and heavy water – made any extensive experimentation program impossible. Nevertheless, at the 30 July meeting of the S-1 Committee, Compton estimated that plutonium would be ready for use in an atomic weapon by fall of 1944.27 The process that appeared to offer the best hope for producing fissionable materials was the electromagnetic method under study at Princeton and at the University of California (Berkeley), where Lawrence’s work with the giant 184-inch magnet attracted the most attention. As with other approaches, this method had been faced with major technical difficulties, but as the months passed Lawrence enthusiastically reported success in meeting and overcoming these problems. Most important, he had actually achieved the separation of small amounts of U-235, even though only in milligram quantities.28
A visit in late July to Berkeley by Colonel Marshall and Stone and Webster representatives convinced them that, in Marshall’s words, “Lawrence’s method is ahead of the other[s] … and should be exploited to the fullest without delay.” The colonel was anxious that work on “a sizeable pilot plant,” as well as a full-scale production plant, begin as soon as possible.29 The S-1 Committee approved
Colonel Marshall’s recommendation on 30 July and decided that the Army, rather than the OSRD, would be responsible for building the pilot plant on land rented from the University of California. Lawrence estimated that material from the electromagnetic process would be ready to go into an atomic weapon by the spring of 1944.
In mid-August, Colonel Nichols visited Berkeley and gave his tentative approval to plans for the pilot plant. With him was Maj. Thomas T. Crenshaw, whose job it was to set up the new California Area Engineers Office of the Manhattan District, to support and assist Lawrence, and to represent Colonel Marshall during construction and operation of the pilot plant. Nichols felt that Lawrence was “making great progress and that the whole project should be pushed into full-scale production as fast as possible,” an opinion with which Lawrence agreed wholeheartedly. Indeed, because Lawrence’s only question concerned the actual efficiency of the separation units, he felt that construction of the full-scale production plant should be started concurrently with that of the pilot installation.30
At the 26 August meeting of the S-1 Committee, both August C. Klein, Stone and Webster’s chief mechanical engineer, and Colonel Marshall supported Lawrence’s proposal, and there was general agreement, based upon Lawrence’s optimistic report, that the electromagnetic method would probably be first to yield material in substantial amounts. The plutonium process, though progressing satisfactorily, was still months away from even the pilot plant stage and the other methods lagged even further behind. Had a decision been made at this time to back a single horse in the nuclear race and to scratch the others, Lawrence very likely would have been the one rider left on the course.
Yet no one was certain that the electromagnetic method would prove to be the best process in the long run. In fact, the group conjectured that the ultimate full-scale plant would probably have several times the capacity of the contemplated electromagnetic production plant and was likely to be comprised of a combination of methods, with one process producing enriched uranium and the electromagnetic method providing the final stage of separation. They thought a decision to proceed with an electromagnetic production plant was unrealistic and might be interpreted as a final decision in favor of the electromagnetic process, causing the development of the other methods to be slowed down – or even eliminated.
At last the conferees at the 26 August meeting agreed to continue work as rapidly as possible on the four pilot plants and on the production of heavy water at Trail. A start on a full-scale plutonium production plant would be delayed, pending the outcome of experiments at the Argonne pilot plant. Design and construction of an electromagnetic production plant would be postponed until mid-September, when the S-1 Committee was to visit the Berkeley project and make further recommendations. Vannevar Bush
approved these conclusions and passed them on to Secretary Stimson with the warning that the time would soon be at hand for a major decision on the extent of the effort the United States should make on the atomic energy program.31