Nuclear Vault Home

Electronic Briefing Books

Special Collection: Key Documents on Nuclear Weapons Policy, 1945-1990

Photo Gallery

Key Readings

 

 

The developments of 1954-1960 that this posting covers have been written up in more detail in an article by William Burr, "The 'Labors of Atlas, Sisyphus, or Hercules'?: U.S. Gas Centrifuge Diplomacy, 1954-1960," in the June 2015 issue of International History Review.

The Gas Centrifuge Secret: Origins of a U.S. Policy of Nuclear Denial, 1954-1960

 

Beginning in 1950s, U.S. Sought to Control Uranium Enrichment Technology that Iranians Are Using Today

In 1954, Washington Ruled Against Brazilian Attempt to Purchase West German Centrifuges for Its Nuclear Program as Contrary to U.S. "Interests"

In 1960, U.S. and Allies Agreed to Put Gas Centrifuge under Secrecy Controls

 

National Security Archive Electronic Briefing Book No. 518

By William Burr

Bill Burr is a senior analyst at the National Security Archive, where he directs the Archive's nuclear history documentation project. See the Archive's Nuclear Vault resources page.

 

For more information, contact:

William Burr at 202/994-7000 or nsarchiv@gwu.edu


  Nixon's Nuclear Specter: The Secret Alert of 1969, Madman Diplomacy, and the Vietnam War

Secretary of State Christian Herter was in the forefront of the effort to classify the gas centrifuge by reaching an understanding with American allies.
(National Archives and Records Administration, Still Pictures Division, RG 59-SO, box 7).
 
  Nixon's Nuclear Specter: The Secret Alert of 1969, Madman Diplomacy, and the Vietnam War

A front-page story in the New York Times partly inspired by leaks of the negotiations over gas centrifuge secrecy. With this article and others that appeared in the international press, the potential of gas centrifuge technology become better known around the world.
Published on 11 October 1960.
 
  Nixon's Nuclear Specter: The Secret Alert of 1969, Madman Diplomacy, and the Vietnam War

Depiction from a 1960 Union Carbide report of the most up-to-date West German gas centrifuge machine.
 
 

Atomic Energy Commission Chairman John A. McCone had initial doubts about classifying the gas centrifuge but became a strong supporter of the diplomatic effort.
 

Washington, D.C., June 29, 2015 – Long before Iran’s nuclear enrichment capabilities – based on gas centrifuge technology – became the center of international negotiations, the U.S. tried to deny that same technology to any country that sought it. In 1954, Washington prohibited a company in occupied Germany from selling gas centrifuges to Brazil, according to declassified documents published today for the first time by the National Security Archive and the Nuclear Proliferation International History Project (NPIHP). These centrifuges had marginal capacity and it would have taken them many years to produce enough highly enriched uranium (HEU) for a nuclear bomb. But – as many in the U.S. government and elsewhere today believe with respect to Iran – the State Department concluded that Brazil’s access to gas centrifuge technology was “contrary to U.S. interests.”

The Brazilian case is an early example of what became a broader U.S. policy of denial of gas centrifuge technology. In 1960, Washington began to consider the technology as a significant nuclear proliferation risk when the latest innovations indicated a significant potential to produce HEU for nuclear weapons on an industrial scale and in secret. According to another document published today, a senior State Department adviser argued that “Should the gas centrifuge process be successfully developed on an unclassified basis it could be utilized in a number of countries either openly or secretly and in either event complicate the problem of preventing the spread of nuclear weapons.”

In 1960, concerns about nuclear proliferation led Secretary of State Christian Herter and Atomic Energy Commission Chairman John McCone to spearhead a little known initiative to put the latest and future gas centrifuge R&D under secrecy and export controls by working with Western European allies who were also developing the technology. Although McCone and others understood that it was difficult to stop scientific-technological advance and that secrecy controls might not work, he believed it was “worth trying,”

 Documents published today by the National Security Archive and the NPIHP shed light on the step-by-step process by which the Eisenhower administration came to be concerned about the potential of the gas centrifuge, reached the conclusion that secrecy and export controls were necessary, and sought an understanding on classification policy with the British, Germans, and Dutch. Such an understanding has in fact been effect since 1960 although it could not stop the infamous A. Q. Khan’s theft of the technology and its spread to Iran and North Korea.

* * * * *

The United States and Gas Centrifuge Technology: The Origins of a Policy of Denial, 1954-1960

By William Burr

Long before Iran’s nuclear enrichment capabilities – based on gas centrifuge technology – became the center of international negotiations, the U.S. tried to deny that same technology to any country that sought it. In June 1954, Robert Terrill, a political officer at the U.S. embassy in Brazil, reported that Brazil’s National Research Council had secretly sent a small group of chemists to occupied West Germany to learn about the workings of gas centrifuge technology for enriching uranium. Committed to a national plan of nuclear energy development, the Brazilians wanted to buy centrifuges from a German firm but U.S. occupation authorities rejected the request, according to declassified documents published today for the first time by the National Security Archive and the Nuclear Proliferation International History Project. These centrifuges had marginal capacity and it would have taken them many years to produce enough highly enriched uranium (HEU) for a nuclear bomb. But the State Department believed that Brazil’s access to gas centrifuge technology was “contrary to U.S. interests.”

The Brazilian case is an early example of what became a broader U.S. policy of denial of gas centrifuge technology. Six years later, in 1960, Washington began to consider the technology as a significant nuclear proliferation risk when the latest innovations indicated a significant potential to produce HEU for nuclear weapons on an industrial scale. Moreover, it could be done in secret. According to another document published today, senior State Department adviser Charles Sullivan argued that “Should the gas centrifuge process be successfully developed on an unclassified basis it could be utilized in a number of countries either openly or secretly and in either event complicate the problem of preventing the spread of nuclear weapons.”

In 1960, concerns about nuclear proliferation led Secretary of State Christian Herter and Atomic Energy Commission Chairman John McCone to spearhead a little known initiative to put the latest and future gas centrifuge R&D under secrecy and export controls by working with Western European allies who were also developing the technology. Although McCone and others understood that it was difficult to stop scientific-technological advance and that secrecy controls might not work, he believed it was “worth trying,” Documents published today by the National Security Archive shed light on the step-by-step process by which the Eisenhower administration came to be concerned about the potential of the gas centrifuge, reached the conclusion that secrecy and export controls were necessary, and sought an understanding on classification policy with the British, Germans, and Dutch. Such an understanding has in fact been effect since 1960 although it could not stop the infamous A. Q. Khan’s theft of the technology and its spread to Pakistan and later to Iran.

Documents in today’s posting include:

  • A telegram from U.S. High Commissioner to Germany James Conant describing a meeting with Brazilian officials and West German scientist Wilhelm Groth to discuss the Brazilian proposal to export centrifuges to Brazil, but with “no (repeat no) bomb development envisaged.”
  • The record of the first briefing, February 1960, by Atomic Energy Commission to CIA, State Department, Defense Department, and White House officials of the potential of new developments in gas centrifuge technology,
  • The AEC’s record of confidential talks in West Germany, the Netherlands, and the United Kingdom during July 1960 of the possibility of applying security classification to the latest developments in gas centrifuge technology.
  • Documents indicating West German concern that security classification would give the appearance of secret research on gas centrifuges for a military program despite earlier anti-nuclear weapons commitments
  • The record of a two hour meeting of the Congressional Joint Committee on Atomic Energy (JCAE) during which McCone, other AEC officials, and Charles Sullivan briefed committee members on the gas centrifuge and its history, proliferation risks, and the problem of secrecy.
  • A January 1967 Atomic Energy Commission task force report on gas centrifuge technology which found that as of late 1966 only 495 Americans had held security clearances for gas centrifuge information compared to 20,000 for gaseous diffusion and over 500,000 for nuclear weapons information.

During the 1930s and 1940s, before top U.S. officials saw the gas centrifuge as a foreign policy issue, scientists in the United States and Western Europe began to appreciate the possibility that it could be used to enrich uranium. If uranium is converted into uranium hexafluoride, which can become a gas at the right temperature, it can flow through high-speed centrifuges for separating the lighter and fissionable uranium isotope, U-235, from the heavier element, U-238. During World War II scientists and technicians working for the Manhattan Project considered the centrifuge as a way to produce highly enriched uranium for a "gun-type" nuclear weapon. Yet project director General Leslie Groves and his advisers stopped the work on centrifuges, in part because the trial models could not operate for long periods of time without breaking down. The Manhattan Project turned to the highly expensive, but more reliable, gaseous diffusion method for producing enriched uranium. [1]

By contrast, German scientists working for the Nazi regime during World War II pursued the centrifuge option as part of their unsuccessful nuclear project. After the war, one of them, Wilhelm Groth, conducted gas centrifuge research at the University of Bonn during the 1950s.[2] As indicated below, when the Brazilians developed their plans for a nuclear power capability, they sought Groth’s aid. During the war, among the prisoners taken by the Soviet Union were the scientists Gernot Zippe and Max Steenbeck; while pursuing gas centrifuge R&D at Soviet installations, Zippe and Steenbeck began solving the technical problems that had stymied the Manhattan Project. They developed relatively small: “subcritical” centrifuges that were easier to operate at high speeds compared to the longer “supercritical” centrifuges that the Dutch later developed. Zippe and other scientists returned to a divided Germany where they continued their research. In an oft-recounted episode, Zippe caught the attention of U.S. intelligence and the Atomic Energy Commission and eventually came to work on an unclassified project at the University of Virginia, enabling him to transfer Soviet technological innovations to the United States.[3]

A posting in 2012 by the National Security Archive featured AEC studies that highlighted concerns about the gas centrifuge and proliferation risk. A February 1960 report by Union Carbide, the contractor that managed Oak Ridge National Laboratory, took Zippe’s research into account. The report, Production of Enriched Uranium for Nuclear Weapons by Nations X, Y, and Z by Means of the Gas Centrifuge Process, had this as a major conclusion:  "in general, it would not be too difficult to build a relatively small clandestine gas centrifuge plant capable of producing sufficient enriched uranium for a small number of nuclear weapons." Industrial countries would have less trouble building and hiding such facilities but, according to Union Carbide, even a nation with "relatively limited industrial activity" could use this technology to produce fissile material for a nuclear weapon in eight years. Moreover, compared to the expensive gaseous diffusion enrichment process, the centrifuge method was relatively cheap – four megawatts would be required for the annual operation of a small plant compared with hundreds, perhaps thousands, of megawatts for a gaseous diffusion plant.

To gauge the proliferation risk, Union Carbide looked at countries at three stages of development. "Class X" highly industrial countries with up-to-date educational systems, such as Germany or Sweden, would need about 5 years to get to the point where they could produce HEU. Class Y" intermediate countries with "limited internal industrial activity," such as Israel or Brazil, would require about 6.5 years to reach the point where they could produce nuclear weapons with the help of gas centrifuges. "Class Z" non-industrial countries with low technological levels, such as Cuba or Egypt, would take about 8 years. They would, however, need the assistance of a "Class X" country and it would not be easy for them to build a completely clandestine plant.

According to Union Carbide, those countries which found a way to produce gas centrifuges for a nuclear weapons program would have to solve other technological problems. Producing HEU for a nuclear weapon would require the creation of a "cascade" of interconnected gas centrifuges and "there has been no indication that anyone has successfully run even two interconnected centrifuges," the report noted. That, however, was a problem of "plant design and process control," which could be solved. Moreover, a high-speed centrifuge required "suitable material" capable of withstanding high velocities. A "class X" country could solve these and other problems within the stipulated eight years.

Officials at the Atomic Energy Commission digested the findings of the Union Carbide study as well as other studies on the competitive potential of the gas centrifuge in producing reactor fuel, such as one by General Electric, but also assessed Zippe’s work at Charlottesville. A major reported prepared in April 1960, "Gas Centrifuge Method of Isotope Separation,"reviewed the latest U.S. research, but also considered the work of the Dutch and the Germans. The belief that the Germans had “the most complete and most extensive gas centrifuge program in the world at this time" reinforced the AEC interest in classification in order to prevent the dissemination of the technology. The AEC also considered the possibility of formal U.S. collaboration with the West German and Dutch centrifuge projects. Although the State Department's reaction to that proposal was equivocal, the Commission staff remained interested. One of their recommendations to the AEC General Manager was to hold "exploratory discussions" with the Dutch and West German governments on the possibility of classifying the latest gas centrifuge technology. Yet, if the Germans and Dutch were resistant to classification, the AEC was willing to consider the possibility of putting the R&D on a non-classified basis.

 As the documents presented below indicate, no one saw classification and export controls as panaceas. Some questioned whether West Germany could even impose secrecy controls without raising suspicions that it was involved in classified nuclear military research in violation of international agreements. British nuclear experts would later argue that with important gas centrifuge research already unclassified it would be possible to build inefficient centrifuges that could produce HEU. That is, secrecy would not eliminate the proliferation risk. More broadly, John McCone wondered whether it would be possible to “sweep [the secret] under the rug.” Noting that industrial countries such as Switzerland could develop gas centrifuge technology if they chose, it “might be an almost entirely impossible thing” to put the secrets “under control.” Yet, McCone believed, “it is worth trying.”

During 1960-1961 the four governments – the British, Dutch, West German, and U.S. – reached what was later known as a “gentlemen’s agreement” to classify gas centrifuge technology. As the documents that follow indicate, the Germans and the Dutch were receptive to U.S. proposals to classify on-going R&D on the gas centrifuge and during the months that followed established government secrecy regimes for the technology. Because of their concern over the political optics, the Germans were insistent upon the confidentiality of the discussions on classification; they were shocked and angry, therefore, when detailed information leaked to a McGraw-Hill trade publication, Nucleonics Today. Nevertheless, the Germans went ahead and, in the face of more leaks in October 1960, publicly announced their decision to classify. The Dutch followed suit in early 1961, although U.S. documents on their decision-making have not surfaced.

Hardly 15 years after the understanding on classification was reached, the determined espionage of Pakistani metallurgist A.Q. Khan thwarted the secrecy controls, later enabling gas centrifuge technology to proliferate further. Although Pakistan had other sources of information about gas centrifuge technology, Khan’s organization played a central role in the efforts to produce cascades of centrifuges for producing weapons-grade enriched uranium. Espionage, however, was not enough to give Pakistan a nuclear capability; Khan and his associates used secret procurement networks to acquire dual-use equipment and technology needed to produce working centrifuges and to build nuclear weapons. During the 1980s and 1990s, the Khan organization expanded its reach by providing Iran with equipment and design information. Apparently the initial transfers were of low quality but transfers during the late 1990s were of advanced, larger “supercritical” centrifuges based on Dutch designs. Whether these transfers gave Iran a meaningful “head start” is a matter of dispute, but the resulting cascades of P-1 centrifuges and their numbers are one of the central issues in the current negotiations over Iran’s nuclear status.[4]

 

I. The Brazilian Case

 

Documents 1A-B: The Secret Team from Brazil

Document A: Letter, Robert P. Terrill, U.S. Embassy, Rio de Janeiro, to Gerard C. Smith, special assistant to the Secretary of State for Atomic Energy, 7 June 1954, Top Secret

Document B: Letter, Terrill to Smith, 11 June 1954, Top Secret

Source: National Archives, Record Group 84, Records of Foreign Service Posts, U.S. Embassy Rio de Janeiro Top Secret General Records, 1944-1955, Box 2, Top Secret 1954

 

Document 2: U.S. High Commissioner telegram 4023 to Department of State, 23 June 1954, Confidential

Source: RG 59, Central Decimal Files 1950-1954, 762A.0221/7-954

 

Document 3: U.S. High Commissioner telegram 75 to State Department, 9 July 1954, Secret

Source: RG 59, Central Decimal Files 1950-1954, 762A.0221/7-954

 

Document 4: State Department telegram 148 to U.S. High Commissioner, 16 July 1954, Secret

Source: RG 59, Central Decimal Files 1950-1954, 762A0221/7-1654

 

Document 5: State Department Instruction CA-2553 to U.S. High Commissioner, Bonn, “Centrifuges for Brazil,” 16 October 1954, Secret

Source: RG 59, Central Decimal Files 1950-1954, 862A.19/10-1654

 

Document 6: U.S. High Commissioner Bonn Despatch 1048 to State Department, “Decisions Taken at the Meeting of the Commissioners of the Military Security Board on November 9, 1954,” 17 November 1954, Secret

Source: RG 59, Central Decimal Files 1950-1954, 862A.19/11-1954

 

Document 7: U.S. High Commissioner Bonn Despatch 1176 to State Department, “Construction of Certain Scientific Research Equipment for the Brazilian Government,” 2 December 1954, Unclassified
Source: RG 59, Central Decimal Files 1950-1954, 862A.19/12-254

 

Years before the 1960 studies that raised the consciousness of top Washington policymakers, U.S. officials had more or less instinctively reached the same conclusion about the risks of freely disseminating gas centrifuge technology when they met Brazil’s attempt to purchase it. For years, political scientists and historians have written about Brazil’s secret attempts, led by Admiral Alvaro Alberto during 1954, to purchase gas centrifuge technology from West Germany so that it could be developed for a Brazilian national nuclear industry. According to one account, the Brazilians sent chemists to West Germany to learn how to handle heavy gases from Wilhelm Groth who negotiated with West German firms to acquire relevant technology for transfer to Brazil. Accounts conflict, however, whether the British or the CIA discovered the operation and halted a secret shipment of centrifuges, but they agree that Washington temporarily put a halt to the attempted transfer of centrifuge technology to Brazil.[5] Much useful information has appeared from Brazilian archives on that country’s nuclear objectives during the 1950s, but nothing yet that sheds light on Admiral Alberto’s mission.

What the U.S. archival record indicates is that Brazil’s attempt to acquire the gas centrifuge involved some subterfuge, but also straightforward, direct approaches to the U.S. government through the Allied High Commission (AHC)—then governing occupied West Germany—including U.S. High Commissioner James B. Conant.[6] As the letter from Robert Terrill indicates [Documents 1A-B], the U.S. government learned that the Brazilians had sent to West Germany a secret team of three chemists to learn about gas centrifuge technology from Groth. What explained the secrecy was recognition that this activity violated Allied occupation statutes for Germany. What the team actually did or whether it was intercepted remains to be learned. But at the same time, the Brazilians worked openly with a German firm, Sartorious-Werke, to buy three low-capacity centrifuges and asked the High Commission’s Military Security Board (MSB) for approval of their manufacture and shipment. [Document 2]. After Admiral Alberto and Groth made an approach to Conant, insisting that the centrifuges were for peaceful uses only, the State Department quickly intervened informing Conant that the transaction contravened AHC law which tightly limited uranium enrichment [See documents 3 and 4]. With West Germany about to be released from Allied occupation authority, the Germans would be free to make their own decisions about centrifuges; in the meantime, however, the State Department believed that the shipment to Brazil was “contrary to U.S. interests” and that the Military Security Board should disapprove it [See document 5].

Records explaining why exactly the State Department opposed the transaction have not yet surfaced publicly, and it is not clear whether the Atomic Energy Commission ever conveyed in its views on the matter. It is likely that the State Department objected to the acquisition of threshold nuclear technology, even if rather basic, by almost any country, in the absence of any international understandings on the peaceful use of nuclear energy (the International Atomic Energy Agency had not yet been created). That Brazil had recently been led by economic nationalists, notably the recently deceased President Getulio Vargas, who were less than responsive to U.S. wishes, may have informed the State Department’s opposition. They may have worried that the nationalists were after the bomb (although Admiral Alberto told Conant that they were not).

Whatever the State Department motives were, on 9 November 1954, the MSB voted against the Sartorius-Werke application, agreeing with the United States that the AHC law should not be amended, a position Washington did not want disclosed to the Brazilians [See document 6]. Shortly thereafter, the Allied High Commission duly informed the Brazilian ambassador to West Germany that the transaction with Sartorius-Werke had been disallowed under AHC law [See document 8]. Nevertheless, as had been anticipated, the Brazilians obtained the centrifuges a few years later from the West Germans. According to the testimony of State Department nuclear expert Charles Sullivan in 1960, those machines were “slightly obsolete” and “I don’t believe there is any particular concern” that the Brazilians had the “capability to develop that on their own to the point where they could produce weapons” [See Document 23].

 

II. The Classification Story

New Developments

 

Document 8: Atomic Energy Commission, “Centrifuge Process for Uranium Enrichment,” 24 February 1960, AEC 610/13, enclosing State Department memorandum of conversation, with same title, 6 February 1960, Secret

Source: Department of Energy mandatory declassification review release

 

Document 9: Letter from Arthur Hartman, Bureau of European Affairs, Office of Atlantic Political and Economic affairs, to Howard Meyers, U.S. Mission to the European Communities, 10 March 1960, Confidential

Source: RG 59, Bureau of European Affairs. Office of Atlantic Political and Economic Affairs, Records Relating to Atomic Energy Matters, 1960-1963, box 2, Centrifuge

 

During February 1960 Union Carbide, an AEC contractor, was completing a study of the proliferation implications of prospective gas centrifuge models: Production of Enriched Uranium for Nuclear Weapons by Nations X, Y, and Z by Means of the Gas Centrifuge Process. At the same time, the AEC was conducting its own studies. Early in the month, senior Commission members met with top officials from the CIA, State Department, Defense Department, and White House to brief them on their preliminary findings. After describing the early development of the technology, AEC research director Paul McDaniel explained that a gas centrifuge plant that could produce 500 kilograms of HEU annually (about 10 nuclear weapons) could be built at relatively low cost and that the technology was within the reach of 20 countries. Yet, this did not mean, AEC Commissioner John F. Floberg later suggested, that all countries that developed the centrifuge wanted the bomb; the possibility that some governments simply wanted to use HEU to fuel nuclear reactors had to be considered.

According to George Kolstad, the director of the AEC’s mathematics division, Washington had several options: 1) to classify work on the technology; 2) to step up U.S. research and development in order to create an on-hand capability to expand reactor fuel production capacity; and 3) to undertake cooperative development work with the Dutch, Germans, and the British. As McDaniel pointed out, the AEC had already made a decision to classify and this had Herter’s support: “keep the work on a classified basis.”

A long chatty letter by Arthur Hartman, then with the Bureau of European Affairs, written a month later, took a critical view of the secrecy option. Stuck at home due to a snow storm, Hartman updated a colleague in Brussels on European Community matters and related issues for U.S. policy. On page 3, in a section that he highlighted for the file copy, Hartman presented one of the findings of the AEC studies: using gas centrifuge technology a “moderately developed country [could] lay their hands on SNM [special nuclear material] at a not prohibitive cost.” Noting that the AEC had raised with the State Department the possibility of reaching agreements with other countries to classify gas centrifuge R&D, Hartman saw that as presenting “terrible problems,” not least because it would raise suspicions that the Germans were conducting secret military nuclear research. Moreover, classification could “hold back development” of the technology and necessitate controls over exports to “friendly” countries. Moreover, German industry, suspicious that classification was a way for Washington to keep its reactor fuel monopoly, could simply “tell both the German [Government] and us to go to hell.” 

As an alternative to national control, Hartman suggested either an internationally-controlled production facility or development under EURATOM auspices, even on a classified basis because the R&D would be associated with peaceful research.

 

 Approaching Bonn

 

Document 10: Memorandum of Conversation, “Manufacture and Export of Ultra-Centrifuge Equipment,” 15 March 1960, Secret

Source: RG 59, Central Decimal Files 1960-1963, 862A.394/3-1560

 

Document 11: Letter, Charles A. Sullivan, Acting Special Assistant to the Secretary of State for Atomic Energy, to Ambassador Walter Dowling, 28 April 1960, enclosing memorandum from Philip J. Farley to Algie Wells, “Control of and Cooperation in Centrifuge Research and Development,” 23 March 1960, attached, Secret

Source: National Archives, Record Group 59, Records of Special Assistant to Secretary of State for Atomic Energy, 1948-1962 [hereinafter SAE], box 298, 12.H Peaceful Uses Subject File.9 Gas Centrifuge 1960-62

 

Document 12: Memorandum of Conversation, “German Development of Ultra-Centrifuge,” 12 May 1960

Source: RG 59, Central Decimal Files 1960-1963, 862A.394/5-1260

 

Document 13: Bonn Embassy Telegram 2312 to State Department, 7 June 1960, Secret

Source: RG 50, Central Decimal Files 1960-1963, 611.62A45/6-760

 

Document 14: Memorandum of Conversation, “German Ultra-Centrifuge Development,” 8 June 1960, Confidential

Source: Source: RG 59, Central Decimal Files 1960-1963, 862A.394/6-860

 

Document 15: Memorandum, Martin J. Hillenbrand, Office of German Affairs, to Foy Kohler, Assistant Secretary of State for European Affairs, “United States-German Cooperation on Control of Ultra-Centrifuge Development,” 13 June 1960, Secret

Source: RG 59, SAE, box 298, 12.H Peaceful Uses Subject File.9 Gas Centrifuge 1960-62

 

With Herter already supporting classification, decisions on gas centrifuge secrecy were above Hartman’s level, but he was correct to see German reactions to any U.S. policy initiative as a central problem. Only a few days after Hartman mailed his letter to Charles Meyer, Herter met with West German Foreign Minister Heinrich von Brentano as part of the preparations for the forthcoming Paris summit meeting, and told him his concerns about the gas centrifuge, especially that West Germany was exporting it as a “normal commercial operation” (possibly a reference to the earlier sales to Brazil). While not familiar with the issue, von Brentano said he would make sure to put a stop to the exports. At the time, the gas centrifuge was often called the “ultra-centrifuge” because of the high speed at which it would be operating.

To prepare the way for eventual talks with the Germans, some weeks later Charles Sullivan, one of Herter’s atomic energy advisers, wrote to the U.S. Ambassador in Bonn, Walter Dowling, to familiarize him with the gas centrifuge problem. The goal would be to reach agreement on export controls, but also to see whether Bonn would be willing to agree to classify West German R&D, which so far had proceeded on an unclassified basis at German universities and corporations. Washington was willing to send AEC and State Department officials to Bonn to discuss the issues. Sullivan’s letter to Dowling included a State Department memorandum by Philip J. Farley which pointed to the problems involving Germany that classification would raise and the possibility of classifying the research under EURATOM auspices.

As the records of conversations with West German officials indicate, they had mixed views on classification and export controls. Following up on the Herter-von Brentano talk, German embassy officials briefed Sullivan on Bonn’s export controls, noting that the centrifuges to Brazil could do very little to produce enriched uranium and assuring him that exports required licenses by the atomic energy and economics ministries. Bonn would be “disinclin[ed]” to approve exports to non-NATO countries. Senior Foreign Ministry official Karl Carstens, however, was skeptical about security classification precisely because it would create suspicions of German weapons research. Indeed, because West Germany was forbidden to possess nuclear weapons as the price for its admittance to the Western European Union (WEU) under the 1954 Treaty of Brussels, he saw the risk of a challenge to the legality of the gas centrifuge research. As a personal suggestion, Carstens thought that the FRG could ask scientists to regulate themselves in order to prevent leaks of sensitive information.

Not long after Secretary of State Herter read the telegram reporting on Carsten’s thinking, he told AEC Chairman McCone that it was a “weak” position because he believed that controls were necessary. Interestingly, McCone said that major U.S. corporations such as General Electric and Dow Chemical wanted to develop the gas centrifuge on an unclassified basis and there was an aversion to the “dead hand of bureaucracy” symbolized by export controls and classification. This worried Herter who saw the technology as a “danger,” but McCone believed that export controls and secrecy could effectively “hold the line” only “for a while.” One long-term solution McCone saw was to saturate the market with cheap U.S. reactor fuel as a way to discourage new entrants into the field (although he did not explain how that would constrain governments with weapons ambitions).

Herter was interested in McCone’s ideas but thought it important to press the Germans to accept classification. However, State Department desk officer for German affairs Martin J. Hillenbrand was not so sure that would work. Disagreeing with McCone’s view that it would take months to “determine the German position,” Hillenbrand believed it could be learned quickly enough once U.S. officials went to Bonn. But he took Carstens’ concerns about the WEU seriously. Another obstacle for the Germans was the industrial firms which could seek compensation for their investments if their work on the gas centrifuge was placed under secrecy controls.

 

The AEC-State Mission

 

Document 16: Philip J. Farley SA/E [Special Assistant for Atomic Energy] to the Acting Secretary, “Gas Centrifuge Discussions with the GFR,” 8 July 1960, with memo from Foy Kohler, 8 July 1960 attached

Source: RG 59, SAE, box 298, 12.H Peaceful Uses Subject File.9 Gas Centrifuge 1960-62

 

Document 17: State Department telegram 81 to U.S. Embassy Bonn, 8 July 1960, Secret

Source: Source: RG 59, Central Decimal Files 1960-1963, 611.62A45/7-760

 

Document 18: Philip J. Farley SA/E to the Acting Secretary, “Briefing Papers on Gas Centrifuge, Communications Satellites, and Safeguards for Indian Power Station,” enclosing paper on “State/AEC Discussions with Germans, Dutch and British on Gas Centrifuge Control Measures,” 22 July 1960, Secret

Source: RG 59, SAE, box 298, 12.H Peaceful Uses Subject File.9 Gas Centrifuge 1960-62

 

Document 19: Atomic Energy Commission, “Report of U.S. Team Concerning the Gas Centrifuge Process in Germany, the Netherlands, and the United Kingdom, July 13-20, 1960,” AEC 610/24, 8 August 1960, Confidential, later regraded to Secret/RD.

Source: Department of Energy mandatory declassification review release

 

Documents 20A-B: Classification Guide

Document A: Department of State Instruction CA-894 to Embassies Bonn, London, and the Hague, “Gas Centrifuge Classification Guide,” 27 July 1960, Confidential

Document B:Department of State Instruction CA-3641 to Embassies Bonn, London, and the Hague,

“Centrifuge Classification Guide – Revision,” 19 October 1960, Confidential

Source: Source: RG 59, Central Decimal Files 1960-1963, 103-AEC/10-1960

Farley and Sullivan started to make arrangements for talks in Bonn and The Hague, but the focus of State Department and AEC thinking was on Germany. The assumption was very likely that if Washington could work out an understanding with Bonn, then the Dutch would probably cooperate. Yet recognizing that a complex internal situation might make the German government unwilling to classify gas centrifuge research and that Washington should not try to “force” Bonn to accept classification, Farley secured Acting Secretary of State Douglas Dillon’s approval for Sullivan and his team to take an exploratory approach.

As it turned out, the meeting with the Germans and the Dutch in mid-July were highly productive, but they also clarified the interdependence of classification decisions. Farley later reported to Dillon that, “somewhat to our surprise,” German officials were responsive to proposals to classify new developments in gas centrifuge technology. The AEC report on the conversations (largely based on telegrams in the National Archive’s 611.62A45 decimal file) indicated that Bonn, like the British and the Dutch, accepted U.S. nonproliferation arguments for classifying gas centrifuge technology. Bonn, however, had a number of concerns, such as the role of EURATOM with respect to classified patents or in the event the FRG made a formal agreement with Washington. Especially worrisome for Bonn was the classification status of gas centrifuge research elsewhere. If British or Dutch gas centrifuge research was unclassified, under relevant law Bonn could not classify its information. A broader concern was that the talks stayed classified; like some of the Americans the Germans worried about the “psychological reactions” to German action to classify gas centrifuge research.

The meeting with Dutch officials produced general agreement on the necessity of classification and the likelihood of a need to consult EURATOM, but the discussion with the scientist Jacob Kistemaker, the leader of Dutch gas centrifuge R&D, illuminated the complexity. The Germans had “widely published” their research on gas flow in centrifuges but also “cascade data” (linkages of centrifuges). In this way they had complicated possibilities of classification, although the fact that (unspecified) technical “know-how” on centrifuges had not been released might be the “key” to restricting information. Kistemaker also worried that West German centrifuge information might have leaked to East Germany, where a leading scientist, Max Steenbeck, had been the “brains” behind the Soviet gas centrifuge program.

In London, Sullivan and his colleagues briefed British officials on the talks with the Dutch and the Germans. Like the Germans, the British were concerned about the public reactions when it became known that Bonn was classifying nuclear research. And they believed that it would be “exceptionally difficult” for the Germans to keep information from their “allies” – possibly a reference to the French. While the British had already classified their gas centrifuge research, they had had questions about the viability of classification when so much information was in the open literature. With the existing information, it might be possible to cascade centrifuges and produce HEU, even if uneconomically. The U.S. team acknowledged this possibility but mentioned Kistemaker’s point about the importance of restricting access to “know-how.” In this connection, the British had the impression that the Brazilians had cascaded their three centrifuges, but agreed with the U.S. team that it was necessary to learn more about the state of Brazilian R&D. 

When the Americans discussed the prospect of classification with the British, Dutch, and Germans they mentioned that the AEC had a classification guide which Washington would share with them. Later in July, the State Department transmitted the guide, which listed major topics that would be classified secret as “restricted data,” such as the design and operation of cascades, of plants, and of “new and improved” centrifuge units, including those that could operate at peripheral speeds of 400 meters per second or more.

 

Leaks

Document 21: State Department telegram 283 to U.S. Embassy Bonn et al., 9 August 1960, Confidential, annotated copy

Source: RG 59, Bureau of European Affairs. Office of Atlantic Political and Economic Affairs, Records Relating to Atomic Energy Matters, 1960-1963, box 2, Centrifuge

Document 22: Bonn Embassy telegram 294 to State Department, 19 August 1960, Secret

Source: RG 59, Central Decimal Files 1960-63, 862A.1901/8-1960.

 

To the chagrin of the AEC, someone leaked an account of “confidential” discussions in West Germany to Nucleonics Week, a McGraw-Hill trade publication. According to the reporter, “the U.S. plea to shroud centrifuge development in West Germany was transmitted at the very highest diplomatic level.” Someone wrote on the reporting telegram: “This is bad. AEC is red-faced about leak. Germans are furious.” That was no understatement: a West German diplomat, Herbert Müller-Roschach, protested: “If this is a sample of what you call classification, I cannot help but believe that our government might be better off without it.”[7] The leaks notwithstanding, the Germans moved forward and implemented a policy to begin classifying the latest gas centrifuge R&D.

 

Briefing Congress

 

Document 23: U.S. Congress, Joint Committee on Atomic Energy, Executive Session, Briefing on Gas Centrifuge Process, 30 August 1960, Secret, excised copy

Source: Record Group 128. Records of the Joint Committees of Congress. Joint Committee on Atomic Energy, 1946-1977, box 45, File 6460: re centrifuge process, 30 August 1960

The Joint Committee on Atomic Energy (JCAE) was a major power in the world of nuclear policy, one that the AEC and other agencies had to keep informed of major developments, such as in gas centrifuge technology. Earlier in the year, AEC Chairman McCone had told the JCAE about the potential of the gas centrifuge for enriching uranium. JCAE Chairman Clinton Anderson asked McCone to return in order to update the committee on recent developments. Accordingly, over the course of a two-hour meeting, McCone reported on matters that he believed were of “of great importance to us and to this country” because of the implications for nuclear proliferation.

With officials from the AEC and the State Department, McCone reviewed the history of gas centrifuge research after World War II, the contribution of Gernot Zippe, the potential of the short-bowl centrifuge, corporate capitalist interest in centrifuge R&D, discussions with the British in late 1959 on the proliferation risk, the decisions to move toward security classification, and the Union Carbide report on the “X” countries. The AEC’s George Kolstad explained how gas centrifuges worked and how a cascade of 10,000 centrifuges could be used to produce low-enriched uranium for fuel or, if reconfigured (“changing the plumbing”), could produce HEU for weapons. Charles Sullivan joined the discussion by reviewing the situation in West Germany and the Netherlands, and the recent exploratory discussions with Dutch and German officials. The hearing concluded with discussion of the leak to Nucleonics, which all agreed was “unfortunate,” especially because it was highly accurate.

The proliferation risk posed by the gas centrifuge was a central issue in the hearing. A statement by Senator Albert Gore, Sr. (D-TN) conveyed the general sentiment: “this gives me an uneasy, uncertain, almost frightening feeling for the world when you consider the irresponsible people who have come to power through coups and otherwise.” Speaking to such concerns, Commissioner Robert E. Wilson suggested that some countries might use a nuclear capability for “blackmail.” He cited Cuba as a potential problem, but also Israel and “any one of a number of small nations that want to be more or less obstreperous and that might, just as France did, make this stupendous effort … to be a member of the [nuclear] Club.” Along the same lines, observing that it “isn’t the Soviets we’re worried about,” McCone cited a motley group – “Japan, Israel, Egypt [and] Argentina” – as potential problem countries, perhaps because he saw them as insufficiently reliable or too volatile. Not seeing “primitive societies” as Nth country risks, McCone jokingly observed that the technology “might very easily fall within the capacity of a group of Chicago gangsters.”

AEC director of classification Charles Marshall saw China at the “forefront” of his concerns, as did Committee member Sen. John Pastore (D-RI), who worried that if China got the technology, “we are in serious trouble.” No one knew that China had already received the technology from Moscow, but China was only one concern in the discussion. Brazil was another. According to Kolstad, because the Brazilians had acquired only three laboratory type centrifuges it would take them “many years” to produce enough fissile material for a weapon. When Marshall assured the committee that the “Brazilians do not know how to make the centrifuges,” Senator Gore probed: “Do you mean Brazil is incapable of reproducing something she has purchased?” Marshall said, yes, because “These are not the easiest things in the world to produce.” The basic problem was a “lack of knowledge. What we hope to accomplish … is that the Germans will stop exporting their knowledge and stop exporting their materials.”

That a leading East German scientist, Max Steenbeck, had helped Moscow develop the gas centrifuge led Senator Anderson to ask if there really was a secret. If the Russians knew it, “how secret is it?” McCone saw that as a fair question and later observed that advanced industrial countries such as Switzerland could get into the centrifuge business if they chose; it “might be an almost entirely impossible thing” to put the secret “under control.” Yet, he believed “it is worth trying.” As the hearings brought out, even if West Germany’s unclassified research was out, their current and future work could be classified and the Dutch were keeping their R&D under wraps. According to Sullivan, “What we are trying to classify are any further developments.” The goal would be to use export and classification controls to put under “immediate control” current research and “the developments that are likely to come about.”

 

German Classification Decision and More Leaks

 

Document 24: Memorandum of Conversation, “Meeting with Professor Groth of the University of Bonn on Gas Centrifuge System, at 1717 H St. (AEC) September 21, 1960,” 21 September 1960, Confidential

Document 25: Memorandum of Conversation, “German Classification of Centrifuge Work,” 4 October 1960, Confidential

Document 26: Memorandum for the Files, Prepared by William Chapin, S/AE, “Press Conference in Bonn on Ultra-Centrifuge Process,” 13 October 1960, Unclassified

Source: RG 59, SAE, box 298, 12.H Peaceful Uses Subject File.9 Gas Centrifuge 1960-62

As the Germans had pointed out in the July conversations with the State-AEC team, classification raised “real internal issues.” For example, when the German scientist Wilhelm Groth (who had worked with the Brazilians some years earlier) visited the AEC in September 1960, he expressed his reservations about “state secrecy,” making it clear that he preferred “industrial secrecy.” Believing it a “bit late” and seeing state secrecy as a disincentive to research and investment, Groth allowed that it could have a “marginal utility” in preventing the spread of nuclear capabilities. This meeting included Jesse Beams, the leader of the gas centrifuge research program at the University of Virginia, who, for security reasons, did not want Groth to pay a visit to Charlottesville.

Despite some objections, Bonn went ahead with a decision to impose “state secrecy” on “sensitivegas centrifuge developments.  When the FRG made it decision, it may have assumed that some level of confidentiality applied, but more leaks dashed any such assumption. Possibly inspired by the Nucleonics Week article, on 11 October 1960 The Washington Post and The New York Times published front-age stories under these headlines: “New Device May Expand Nuclear Club” and “Cheaper A-Bomb Process is Threat to Arms Control,” with more following: “U.S. Asks Secrecy on `Cheap’ A-Bomb,” and “Bonn Classifies Atomic Process.” The story was out; probably besieged by questions, the German Foreign Ministry held a process conference on 13 October which confirmed the decision to classify. But ignoring its own positive interest in classification, Bonn blamed Washington: “the action was taken as a result of American suggestions.”

The decision-making in the Netherlands was more protracted; it was not until early 1961 that the Dutch made their decision to set up a government-controlled classification system.[8] That development was the last step in the making of what became known as the “gentleman’s agreement” on gas centrifuge secrecy. The use of that gendered language reflects upon the times, in terms of who was making and implementing policy, but it also brings to mind the informal, unwritten nature of the understanding that Washington and close allies reached to try to prevent leaks in the latest developments of a sensitive technology.

 

Document 27: A. E. Cameron, Oak Ridge National Laboratory, “Assessment of a Control System [:] Detection of Hidden Isotope Separation Facilities,” 10 April 1961, Secret, excised copy

Source: Department of Energy mandatory declassification review release.

A report prepared by A.E. Cameron, a chemist at Oak Ridge Laboratory, in early 1961 underlined the problem of detecting covert gas centrifuge plants. The report was commissioned as background for consideration of a fissile material production cut-off that was under discussion as part of a general nuclear disarmament agreement: A key problem in enforcing a cut-off was how would “undeclared” or “clandestine” fissile material production facilities be detected? In this report, various methods of isotope separation were considered, but also uranium metal production plants and plants for producing other important isotopes. Cameron saw the gaseous diffusion method as presenting tough detection problems under some circumstances, but gas centrifuge facilities could be especially tough targets for detection: “the chances of locating a centrifuge plant are very poor if the facility was constructed with the intention of clandestine operation.” Making detection problematic were such factors as relatively low power consumption, the small size of a plant, and a small uranium inventory.

The report briefly reviewed the electromagnetic process for isotope separation, noting its important role in producing HEU for the “first [atomic] bomb used in warfare,” but wrote it off as a proliferation risk: the “method is not appealing enough to be seriously considered as a means of producing appreciable amounts of U235 for weapons purposes under present technology.” While the Oak Ridge scientists dismissed the electromagnetic process as inefficient and overly labor intensive, during the 1980s and early 1990s, it was Saddam Hussein’s method of choice for producing fissile material for Iraq’s nuclear weapons program.[9]

 

Document 28: U.S. Atomic Energy Commission, Task Force Report, A study of Gas Centrifuge as it Relates to the Proliferation of Nuclear Weapons, n.d. [circa January 1967], Top Secret, Excised copy

Source: Department of Energy mandatory declassification review release, under appeal

Prepared at the request of the AEC just as the U.S. government was putting the negotiation of a nuclear nonproliferation treaty on its agenda, this highly classified report showed continued support for the multi-country gas centrifuge classification system that had developed in the early 1960s. After reviewing major developments in gas centrifuge R&D, which apparently included significant but highly secret technological advances, the report considered the potential contribution of the gas centrifuge to the nuclear proliferation problem in some detail. After reviewing possible motives for proliferation, the AEC Task Force found that “The gas centrifuge, especially with the AEC’s current advanced technology, could serve both to increase the number of Nth powers and to accelerate the rate of entry into the ‘club.’” While the gas centrifuge was not necessary for nuclear proliferation, it had enough advantages to make it a technology of choice. The Task Force supported the existing classification restrictions and export controls, including the arrangements with the Dutch and Germans. But the report acknowledged that even if a would-be proliferant could not get the latest technology, the possibility of making “significant progress” on the basis of unclassified information could not be ruled out.

The report has noteworthy features, including statistics [see p. 60 of PDF] on how many U.S. citizens held security clearances for gas centrifuge technology compared to gaseous diffusion and in overall comparison to nuclear weapons technology. Through 1966, only 600 government and corporate employees had had clearances for gas centrifuge technology, compared to 20,000 for gaseous diffusion, and somewhere in the range of a half-a-million for nuclear weapons secrets.

* * *


[1]. For useful background, see R. Scott Kemp, "The End of Manhattan: How the Gas Centrifuge Changed the Quest for Nuclear Weapons," Technology and Culture 53 (2012): 276-279.

[2]. For Groth, see Jeremy Bernstein, Hitler’s Uranium Club: The Secret Recordings at Farm Hall (New York: Copernicus Books, 2001), 1; Samuel A. Goudsmit, Alsos (Woodbury, NY: AIP Press, 1996), 94-95.

[3]. Kemp, “The End of Manhattan,” 279-297. For background on Zippe, see also presentation by Houston Wood, University of Virginia, at Woodrow Wilson Center for International Scholars, 20 January 2010.

[4]. Feroz Khan writes that Khan’s assistance gave Iran a “head start,” but Kemp argues that Iran’s centrifuge program could have developed more effectively by using the short “subcritical” centrifuges of the type that the Soviets had developed and Gernot Zippe had advanced. Feroz Hassan Khan, Eating Grass: The Making of the Pakistani Bomb (Stanford: Stanford University Press, 2012), 373; R. Scott Kemp, “The Nonproliferation Emperor Has No Clothes: The Gas Centrifuge, Supply-Side Controls, and the Future of Nuclear Proliferation,” International Security 38 (Spring 2014), 70.

[5]. See Norman Gall, “Atoms for Brazil, Dangers for All,” Foreign Policy No. 23 (Summer 1976): 181-182; Jean Krasno, “Brazil’s Secret Nuclear Program,” Orbis 38 (1994), 429-430, and John R. Redick, Nuclear Illusions: Argentina and Brazil (Washington, D.C.: Henry L. Stimson Center, 1995), 6. For background on the Brazilian nuclear program, see Carlo Patti, “Origins and Evolution of the Brazilian Nuclear Program (1947-2011),” n.d., Nuclear Proliferation International History Project.

[6]. For background on Conant and his role as High Commissioner to Germany, see James G. Hershberg, James B. Conant: Harvard to Hiroshima and the Making of the Nuclear Age (New York: Knopt, 1993).

[7]. Later in the 1960s Müller-Roschach was retired from the diplomatic service when it was learned that after World War II he had changed his name from Herbert Müller to conceal his wartime role in the Foreign Office’s Jewish Affairs Department (Judenreferat). See Christopher Browning, “The Foreign Office Revisited,” Bulletin of the German Historical Institute, Fall 2011, 75.

[8]. Abel Streefland, Leiden Observatory, “Putting a Lid on the Gas Centrifuge: Classification of the Dutch Ultracentrifuge Project, 1960-1961,” work in progress.

[9]. David Albright and Mark Hibbs, “Iraq's Bomb: Blueprints and Artifacts,” The Bulletin of the Atomic Scientists 48 (January 1992) 30-40.