Content uploaded by R. Scott Kemp
Author content
All content in this area was uploaded by R. Scott Kemp on Mar 18, 2016
Content may be subject to copyright.
272
Introduction
The first nuclear weapons were born from technologies of superindus-
trial scale. The Manhattan Project exceeded the domestic automobile in-
dustry in its size. The gaseous-diffusion plant that enriched uranium at the
Oak Ridge National Laboratory in Tennessee employed at its peak some
12,000 people, enclosed forty-four acres under a single roof, and by 1945
consumed nearly three times the electricity of the highly industrialized city
of Detroit.1In the 1940s and ’50s the making of nuclear bombs was under-
stood to be a massive undertaking that required vast resources and nearly
unparalleled human ingenuity. The U.S. atomic enterprise encouraged a way
of thinking about nuclear proliferation that was intimately tied to technol-
ogy and industry. In the words of President Harry S. Truman, it seemed
“doubtful if such another combination could be got together in the world.”2
The difficulty was not in the bomb per se—scientists had warned that
this step would not be hard to replicate—but rather in the apparently mas-
sive effort needed to produce the nuclear-explosive materials that fueled the
R. Scott Kemp studies problems of international security by combining physics, history,
and public policy. He received his Ph.D. from Princeton University and is currently an
associate research scholar at Princeton’s Program on Science and Global Security. In
completing this article he is indebted to Michael Gordin and the editors and reviewers
of Technology and Culture for their input.
©2012 by the Society for the History of Technology. All rights reserved.
0040-165X/12/5302-0002/272–305
1. U.S. Atomic Energy Commission (AEC), AEC Handbook on Oak Ridge.
2. Harry S. Truman, “Statement by the President Announcing the Use of the A-
Bomb at Hiroshima”; Richard Rhodes, The Making of the Atomic Bomb; Lillian Hodde-
son, Paul W. Henriksen, Roger A. Meade, and Catherine L. Westfall, Critical Assembly;
Vincent C. Jones, Manhattan, the Army and the Atomic Bomb; Cynthia C. Kelly, ed., The
Manhattan Project.
The End of Manhattan
How the Gas Centrifuge Changed the Quest
for Nuclear Weapons
R. SCOTT KEMP
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 272
KEMPK|KGas Centrifuges and Nuclear Weapons
273
bomb.3General Leslie Groves thought such an effort would take the Soviets
“fifteen to twenty years—more likely the latter.”4Specifically,Groves felt the
greatest secret of the bomb was in the industrial organization and tech-
niques required, but even these he felt would be developed by the Soviets
given sufficient time.5Truman was so convinced of Soviet backwardness
that, upon learning about their first nuclear test, he refused for a time to
believe it to be true.6Truman and Groves were not the only ones to hold
this view. Treasury Secretary John Snyder, U.S. Ambassador to the Soviet
Union Walter Bedell Smith, Secretary of Defense James Forrestal, and
numerous other Kremlinologists all agreed, as Forrestal put it, that “[t]he
Soviet Union could not possibly have the industrial competence to make
the atomic bomb.”7While others, such as Secretary of War Henry Stimson
and Secretary of State Dean Acheson, were more skeptical that the United
States could maintain an “atomic monopoly,” the apparent failure of the
Baruch Plan to bring about a system of international control left, as Tru-
man put it, “no alternative [but] to maintain, if we could, our initial supe-
riority in the atomic field.”8
The mythology of atomic-industrial superiority was codified in the
Atomic Energy Act of 1946, which outlined a system of secrecy and tech-
nology control as the primary mechanisms for preventing nuclear prolifer-
ation—a system that could only work to the extent that the myth was true.9
When the Soviet Union acquired weapons in 1949 and the United King-
dom in 1952, both disruptions could be eased into the mythological frame-
work without shattering it: namely,that the Soviets had been advantaged by
espionage, and that the British were collaborators on the Manhattan
Project, indeed had founded it, and had come to learn the secrets alongside
American scientists.10 Until the French nuclear weapons test and perhaps
3. James Franck et al., “Report of the Committee on Political and Social Problems”
(Franck Report).
4. Michael D. Gordin, Red Cloud at Dawn, 69–70; Nuclear Task Force, “Nuclear In-
spection.”
5. Alex Wellerstein, “Knowledge and the Bomb,” 173.
6. Secretary of Defense Louis A. Johnson also believed it was a reactor accident; see
Gordin, Red Cloud at Dawn, 219ff.
7. Ibid., 70–71 (emphasis added). As an example of the hubris that surrounded the
Manhattan Project, Groves, in trying to refine his estimate, telephoned G. M. Read of
DuPont, which had built the plutonium-production facility at the Hanford Nuclear Res-
ervation in Washington State. Read informed Groves that “[e]ven if they had all the
plans, I don’t think they would live long enough to build one of these things.”
8. Lawrence S. Wittner, The Struggle Against the Bomb, 247.
9. For an excellent account of how the system of secrecy was established, see Weller-
stein, “Knowledge and the Bomb.”
10. Academic historians generally consider that spies only helped to speed the Soviet
program by a modest amount, but that they were not crucial; see, for example, Gordin,
Red Cloud at Dawn.
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 273
TECHNOLOGY AND CULTURE
APRIL
2012
VOL. 53
274
even after it, it would have been feasible to believe that nuclear weapons
were a privilege of the technological elite.11
There is, however, one technology for which the story is exactly re-
versed: the uranium-enrichment gas centrifuge. This machine was created
in the Soviet Union and conveyed to the United States by spies and inform-
ants. Unlike any of its predecessor technologies, it was small, inexpensive,
and relatively simple to make, yet the gas centrifuge was just as capable of
enriching uranium for nuclear weapons. Today, it is best known as the de-
vice that made nuclear weapons available to the developing world, but ac-
counts of its proliferation tend to highlight the machinations of a murky
nuclear black market and leave the centrifuge itself as a mysterious, some-
times glorified technical gem.12 The existing literature lacks an adequate ex-
planation of how and when this device came to change the dynamics of
nuclear proliferation.
Several historical aspects of the centrifuge are well known. For example,
it is widely reported that the Manhattan Project was not successful in devel-
oping centrifuges, but that after World War II, German prisoners of war
helped to perfect the centrifuge in a Soviet labor camp. Little has been writ-
ten on what was required or the consequences these developments had for
nuclear proliferation. This article traces the centrifuge’s development and
explores the role of technological change and tacit knowledge in the on-
ward proliferation of the device. The findings presented here show that the
centrifuge was never a sophisticated or resource-intensive technology,but a
rather simple one that only became simpler over time. It is a machine that
breaks with the Manhattan Project’s legend of techno-industrial greatness
and invalidates the technology-based nonproliferation controls that flowed
from it.
Recently released intelligence reports and memoirs have provided new
information about centrifuge development in the Soviet Union. Notably,
participants have acknowledged that Soviet contributions were of equal
importance to those made by German scientists. Also used here are the per-
sonal papers of the former director of the U.S. centrifuge program, Ralph
Lowry, which were officially declassified by the Department of Energy in
1985 though never released to the public. The archive contains over thirty-
three linear feet of records and was stored in Lowry’s office until he died in
11. This was especially the case with enrichment. Myron Kratzer, who served as head
of the AEC Division of International Affairs, reports that “[i]n those days [1945–70] we
really did believe we [the United States] were the masters of enrichment and that nobody
could compete with us in the field.” He added: “We did worry about the centrifuge in the
1960s, however” (Kratzer, personal interview with author).
12. See, for example, the following popular texts: Catherine Collins and Douglas
Frantz, Fallout; David Albright, Peddling Peril; Frantz and Collins, The Nuclear Jihadist;
William Langewiesche, The Atomic Bazaar; Gordon Corera, Shopping for Bombs; and
Mahdi Obeidi and Kurt Pitzer, The Bomb in My Garden.
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 274
KEMPK|KGas Centrifuges and Nuclear Weapons
275
2007. This article is the first to make use of the Lowry archive, which de-
scribes a remarkably small-scale development effort. This scale directly
contradicts the image of proliferation as a massive undertaking during the
1940s and ’50s.
In analyzing both the early Soviet and U.S. programs, this article argues
that a centrifuge was technically feasible during the Manhattan Project, but
was not successfully developed.The Manhattan Project’s centrifuge had sev-
eral critical shortcomings, and the project was shut down before its devel-
opers had an opportunity to resolve them. Furthermore, Soviet spies helped
the USSR to abandon a failing enrichment effort at an early date, and thus
likely accelerated the pace of the uranium side of the Soviet bomb effort,
perhaps substantially. Third, when later in the Soviet program a German
team finally did solve the riddle of the centrifuge, the result was a device of
incredible simplicity; so basic were its engineering requirements that only a
minimal staff and no precision engineering were required. And fourth, that
the tacit knowledge transferred from the Soviet Union to the United States,
and in turn from the United States to others, was insignificant. As such, this
kind of knowledge appears not to be important in the replication of cen-
trifuges by other countries.
Taken together, these findings suggest that the barriers to the onward
proliferation of the centrifuge were not technical in nature. This did not,
however, change the course of nonproliferation policy, which continued to
focus heavily on technology control. Because of centrifuges states were able
to pursue nuclear weapon capabilities despite technology-control regimes,
as China did with its centrifuge program that began in 1957 and remained
unknown to U.S. officials for decades.13 While there were efforts in the
1960s to move past technology controls and reduce the demand for nuclear
weapons among the developed nations of Europe by establishing the Multi-
lateral Force within the NATO security coalition, the problem was, in fact,
more global. Technology restraints were no longer serious barriers to pro-
liferation in Europe or among many of the developing nations; for exam-
ple, by 1975 countries like India, Brazil, and Pakistan also had centrifuge
programs. Yet nearly every other nonproliferation institution created at the
initiative of the United States has been designed to restrict access to nuclear
technology, not to reduce demand.14 Only the Treaty on the Non-Prolifera-
tion of Nuclear Weapons (initiated in 1968; in force since 1970) had the
potential to reduce demand in a universal way, and yet even it became lit-
13. Dangdai Zhongguo, ed.,“Research on Nuclear Science and Technology, Section 5”;
Central Intelligence Agency (CIA),“Communist China’s Advanced Weapons Program.”
14. These include the Zangger Committee, the Nuclear Suppliers Group, U.N. Reso-
lution 1540, the Proliferation Security Initiative, and, most recently, new proposals like
fuel-banks, multilateral fuel-cycle regimes, and cradle-to-grave nuclear-energy frame-
works. For a history of counterproliferation thought in the United States,see Joseph Pilat
and Walter Kirchner, “The Technological Promise of Counterproliferation.”
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 275
TECHNOLOGY AND CULTURE
APRIL
2012
VOL. 53
276
tle more than a mechanism for mandating safeguards, just another kind of
technology control.
A Machine Is Born
Although in January 1919 the wounds of World War I were still fresh
and nations were racked by debt, life in the field of physics was exciting.
The existence of isotopes had been discovered just six years earlier: a new
mystery of matter, in which chemically identical atoms of the same element
appeared to have different masses and radioactive properties. Because they
are chemically identical, the only known way to separate one isotope from
another was by the minute differences in their masses. The question for ex-
perimental physicists was how to go about sorting isotopes, atom by atom,
when presented with a bulk quantity of matter. As President Woodrow
Wilson was sailing to Europe for the Paris Peace Conference, British Nobel
laureate Frederick Lindemann along with the inventor of the mass spectro-
graph, Francis Aston, mused about how they might attempt to produce a
quantity of pure isotopes. They wrote that “[n]one of the physical methods
considered give hope of easy separation. The most promising method ap-
pears to be the use of a centrifuge, provided the engineering problems can
be overcome.”15
Throughout the 1920s the engineering problems proved formidable,
with scientists struggling for more than a decade to build a working proto-
type. Notables like Robert Mulliken in the United States, Paul Harteck in
Germany, and Sydney Chapman in the United Kingdom were each unsuc-
cessful in their attempts.16 The problem was that their centrifuges could not
spin fast enough to effect a measurable separation; when higher speeds
were finally achieved, air friction heated and convectively mixed the iso-
topic gases, thus canceling any separative effect. In 1934 Jesse Beams of the
University of Virginia had the insight to build a centrifuge inside a vacuum
chamber.17 The vacuum removed air friction from the equation and ther-
mally isolated the process gas from the temperature fluctuations of the out-
side world. With this idea Beams was able to demonstrate the first centrif-
ugal separation of isotopes, separating chlorine-35 from chlorine-37.
15. Frederick A. Lindemann and Francis W. Aston, “The Possibility of Separating
Isotopes,” 523–34.
16. Robert S. Mulliken, “The Separation of Isotopes by Thermal and Pressure Diffu-
sion,” 1033ff, and “The Separation of Isotopes,” 1592ff; K. Beyerle, P. Harteck, W. Groth,
and H. Jensen, Über Gaszentrifugen.
17. J. W. Beams and F. B. Haynes, “The Separation of Isotopes by Centrifuging,”
491ff; Beams, Early History of the Gas Centrifuge Work in the U.S.A.
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 276
KEMPK|KGas Centrifuges and Nuclear Weapons
277
Not Ready for Prime Time
When the question of a uranium bomb arose in the spring of 1940, it
seemed that a significant quantity of enriched uranium might be needed,
and from the beginning the centrifuge was a favored method.18 According
to Richard Hewlett and Oscar Anderson, many of the key figures subse-
quently involved in establishing the research program had, just a few weeks
earlier, attended a meeting of the American Physical Society, in which the
view emerged that the centrifuge was the only process that seemed to offer
much hope of large-scale enrichment.19 Of those present at the meeting,
only two persons—Beams and Alfred Nier—had firsthand knowledge of
isotope separation. Although experiments on uranium-235 had been done
using small samples prepared by Nier in his mass spectrographs, this elec-
tromagnetic method did not appear particularly scalable. Beams, on the
other hand, had never separated uranium, but he had been thinking about
it since March 1939 and was apparently persuasive in describing the poten-
tial of his centrifuge as suitable for uranium enrichment on a large scale.20
A research program was funded under the auspices of the Office of Sci-
entific Research and Development (OSRD) in August 1941. While multiple
methods were investigated the largest share of funds went to Beams and his
centrifuges—nearly four times the amount allocated to gaseous diffusion,
the next best-funded technology. A short (read: practical) centrifuge was to
be designed at Columbia University and built by Westinghouse Research
Laboratories; a longer, high-performance model was to be developed by
Beams himself at the University of Virginia.21 Conceptually, both designs
were derived from the machine Beams had made as a laboratory apparatus
to separate chlorine isotopes. The first Columbia model was completed in
18. Enriched uranium was thought to be necessary to demonstrate the fission chain
reaction if Leo Szilard and Enrico Fermi’s carbon-moderated pile did not work, and it
was thought almost certainly necessary if the bomb required fast neutrons, as many sus-
pected it might; see Richard G. Hewlett and Oscar E. Anderson, The New World, 1:22.
19. Ibid., 1:22–23.
20. A third attendee, Ross Gunn, who was a technical advisor to the Naval Research
Laboratory, was likely familiar with the work of Philip Abelson on thermal diffusion,
which had been ongoing at the laboratory. The process is not energetically favorable, as
had already been demonstrated by W. H. Furry, R. Clark Jones, and L. Onsager during
the previous year (see Furry, Clark Jones, and Onsager,“On the Theory of Isotope Sep-
aration by Thermal Diffusion”). Beams documented his early thinking in Beams, A. C.
Hagg, and E. V. Murphree, Developments in the Centrifuge Separation Project, 26.
21. This was essentially a continuation of his current research. Beams did, however,
use his new funds to test the feasibility of using uranium hexafluoride (UF6), but this he
did using an even more primitive centrifuge concept called the “evaporative centrifuge,”
which was single-batch process and nothing like the machines he intended to be built for
the Manhattan Project. In 1940 most of the funds were used to hire more graduate stu-
dents, who worked on experiments using the batch machine—arguably a poor use of
wartime funding (see ibid., 27).
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 277
TECHNOLOGY AND CULTURE
APRIL
2012
VOL. 53
278
December 1941 and construction of a pilot plant at the Standard Oil Devel-
opment Company commenced, anticipating the imminent delivery of
working centrifuges. However, as the year progressed, tests on the prototype
machines built by Westinghouse and the University of Virginia showed per-
sistent problems.
The spinning tube of the Beams centrifuge was coaxially mounted be-
tween two drive shafts. The tube was surrounded by the critical vacuum
casing, and the drive shafts protruded through openings in the casing,
which were made vacuum-tight with the aid of heavy-grease seals. The
ends of the drive shafts were then supported outside of the case with con-
ventional rigid journal bearings.22 At high speeds friction in the bearings
and vacuum seal resulted in a tremendous amount of energy lost as heat—
approximately a thousand watts, compared to the loss of about only one
watt in a modern centrifuge. Besides being an energy sink, this friction also
caused the components to wear out quickly. Such a machine was capable of
short runs, as required in a lab environment, but would not stand up to the
demands of industrial-scale operation. Attempts were made at reducing the
friction by using air bearings, but no rethinking of the basic design ever
occurred.
Throughout 1942 it was the view of the S-1 Uranium Committee, the
predecessor of the Manhattan Project, that all enrichment technologies
should be pursued simultaneously. As time passed it had become increas-
ingly apparent that the favored centrifuge was not going to meet expecta-
tions. By November, plans for the Standard Oil pilot plant were placed on
hold, and on 26 October James Conant reported to OSRD director Vanne-
var Bush that while no single enrichment process had emerged as superior,
the centrifuge was definitely the weakest.23 According to Hewlett and An-
derson, Conant and Manhattan Project director General Leslie Groves were
looking to accelerate the uranium-enrichment program, which would re-
quire them to make financial tradeoffs among the competing methods. At
a meeting of 10 November the two decided to pursue a full-scale gaseous-
diffusion plant, with the electromagnetic method held in reserve and, as
such, the centrifuge could be dropped.24 Hewlett and Anderson’s docu-
mentation of the centrifuge ends there, but Cameron Reed has shown that
the centrifuge enjoyed a brief revival a few months later. Around January
1943 engineering studies for the gaseous-diffusion plant revealed that the
diffusion membranes would be efficient only up to about a 36.6 percent
uranium-235 concentration.25 Because of the nonlinear nature of enrich-
22. These bearings were rigid, in that they did not allow the centrifuge rotor to pivot
like a top, as do modern centrifuge bearings. The design is detailed in ibid., 141–42.
23. B. Cameron Reed, “Centrifugation during the Manhattan Project,” 433; Hewlett
and Anderson, The New World, 102.
24. Hewlett and Anderson, The New World, 107–8.
25. Reed,“Centrifugation during the Manhattan Project,” 434.
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 278
KEMPK|KGas Centrifuges and Nuclear Weapons
279
ment, this represented about 96 percent of the total effort required, but it
still fell well short of being weapons-grade. Therefore another technology
would be needed.
In February 1943 Eger Murphree of the Standard Oil Development
Company, his firm having lost its contract for the previously planned pilot
plant, pitched the idea for a much smaller “topping-off ” plant, in which
centrifuges would be used to enrich uranium from 36.6 to 90 percent. No
new action was taken at the time of the proposal, and over the course of
1943 work on individual centrifuges continued, but the friction and relia-
bility problems inherent in the design persisted. The S-1 Uranium Com-
mittee met in September 1943 and agreed that existing centrifuge research
should continue at as small an expense as possible, with single-machine
tests at Standard Oil, but that Murphree’s proposal for a topping-off plant
would not be pursued (an electromagnetic isotope-separation proposal
would win this contract, leading to the construction of the Y-12 plant at
Oak Ridge).26
As December approached and research contracts were due to expire
there was much lobbying of S-1 committee members by Murphree. Groves
had to act. He solicited the opinions of the entire S-1 Uranium Committee.
Reed reports that views on the prospects for the centrifuge were mixed,
though mainly negative. On 19 January 1944 Groves indicated in a letter to
Conant and copied to the entire S-1 Uranium Committee that centrifuge
contracts would be allowed to expire at the end of the month, without ex-
tension.27 This was just as well, because four days later a nut on the Stan-
dard Oil centrifuge connecting the lower driveshaft to the rotor gave way,
resulting in the destruction of the test device. The machine had lasted only
ninety-nine days.28
Not Ready for Прайм-Тайм
In the Soviet Union there had also been a long-standing interest in iso-
tope separation for research purposes, with a view toward harnessing
atomic energy. The two most favored technologies in 1940–41 were ther-
mal diffusion and the gas centrifuge, although other ideas were being pur-
sued.29 Thermal diffusion had the greater initial research base, with work
being done at the Biogeological Laboratory and the Dnepropetrovsk Phys-
26. Ibid., 438. For the political context of these decisions, see Jones, Manhattan, the
Army and the Atomic Bomb, 243–44.
27. Reed,“Centrifugation during the Manhattan Project,” 439.
28. Beams, Hagg, and Murphree, Developments in the Centrifuge Separation Project,
196.
29. Notably, Artsimovich pursued electromagnetic isotope separation at Ioffe’s insti-
tute, and linear accelerators were being tried at the Radium Institute; see David Hollo-
way, Stalin and the Bomb, 11.
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 279
TECHNOLOGY AND CULTURE
APRIL
2012
VOL. 53
280
30. An early rotor was reported to be twenty-five centimeters in diameter and sixty-
centimeters long, with a half-centimeter-thick wall; it sat on a seven-centimeter-diame-
ter shaft mounted across three ball-bearing supports. Internally, the rotor was divided by
discs into twenty, forty, or sixty separation chambers depending on the configuration. In
total, the drive, support mechanism, and rotor weighed about two-and-a-half tons and
required three kilowatts of power to operate. Depending on the voltage, the rotor’s max-
imum speed was in the range of 105 to 130 meters per second—far below what is neces-
sary to effect a measurable separation. See A. A. Sazykin, “Development of Gas Centri-
fuges for Uranium Enrichment in the USSR”; see also the participant history of D. L.
Simonenko, “Kratkoe opisanie pervykh eksperimental’nykh rabot po razdeleniyu izo-
topov v SSSR.”
31. Holloway, Stalin and the Bomb, 77.
32. For an explanation of gravity modes, see John M. Vance, Rotordynamics of Tur-
bomachinery, 27; and Agnieszka Muszyn´ ska, Rotordynamics, 171. For Lange’s problems,
see Holloway, Stalin and the Bomb, 99.
33. Sazykin, “Development of Gas Centrifuges for Uranium Enrichment in the
USSR”; Simonenko, “Kratkoe opisanie pervykh eksperimental’nykh rabot po razde-
leniyu izotopov v SSSR”; Holloway, Stalin and the Bomb, 99.
iochemical Institute, but the centrifuge would gather supporters. David
Holloway argues that a number of physicists expected that thermal diffu-
sion would use more energy than could be extracted from the resulting
purified uranium-235 and thus viewed it as unsuitable for the basic mis-
sion of atomic power—in fact, they were right. One of these was V. S. Shpi-
nel’ of the Ukrainian Physicotechnical Institute (UFTI), who supported the
Soviet Union’s only centrifuge program, which was run by German émigré
Fritz Lange. Lange’s centrifuges were heavy, noisy, and clumsy machines
that operated horizontally on roller bearings and produced no measurable
enrichment.30 In the autumn of 1941 Georgii Flerov, prompted by his sus-
picions that the United States was building a nuclear bomb, began to ex-
plore what would be required. He worked out that a fast-fission chain reac-
tion was necessary, and that this would require enriched uranium.31 Thus
when later intelligence about the MAUD (Military Application of Uranium
Detonation) Committee—a group organized by the British government to
investigate the possibility of a nuclear bomb—bootstrapped the Soviet
weapon program, and Igor Kurchatov’s Laboratory No.2 was established in
April 1942, Lange’s centrifuge work became part of it. Presumably under
pressure to show results, Lange attempted to increase the device’s perform-
ance by building a longer centrifuge (now five meters in length), which only
exacerbated the engineering problems inherent in his design. Horizontal
centrifuges do not work because the tubes sag in the middle from gravity.
This causes destructive vibrations when they are operated at speed, as con-
firmed in Holloway’s account of Lange’s program.32 While Beams’s cen-
trifuge had lasted only ninety-nine days, the bearings in Lange’s machines
were capable of only eight to ten hours, and if pushed to high speeds, his
longer machines would snap in the middle.33
According to Holloway, Kurchatov was shown the MAUD Report and
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 280
KEMPK|KGas Centrifuges and Nuclear Weapons
281
34. Holloway, Stalin and the Bomb, 91.
35. Ibid.
36. Ibid., 103–5.
37. Ibid., 97.
38. Ibid., 99. The date of Lange’s transfer is from Arkadii Kruglov, The History of the
Soviet Atomic Industry, 130.
39. Holloway, Stalin and the Bomb, 105; Alexei B. Kojevnikov, Stalin’s Great Science,
140.
40. Beams, “High Rotational Speeds” and “High Speed Centrifuging.”
41. Lange was not alone in favoring horizontal machines; Jacob Kistemaker in the
Netherlands also pursued horizontal machines during the mid-1950s.
other collected Soviet intelligence in March 1943. He noted that the MAUD
Committee had strongly favored gaseous diffusion over other methods,
and other intelligence confirmed the flaws of thermal diffusion and pro-
vided information about British diffusion efforts. Kurchatov was surprised
to learn about the prospects of gaseous diffusion and remarked that, with
this new information, it would now be possible“to begin here in the Union
a new and very important direction for resolving the problem of isotope
separation.”34 He further noted Britain’s apparent lack of interest in the
centrifuge, but was not willing to abandon the method until Lange’s ma-
chine had been given a fair trial.35
By the beginning of 1944 Kurchatov learned from intelligence that the
United States had chosen diffusion over centrifuges, and he was also receiv-
ing information from Klaus Fuchs, who was now in the United States work-
ing on diffusion membranes.36 Kurchatov had put Isaak Kikoin in charge of
isotope separation, but could not reveal any intelligence to him directly. Hol-
loway notes that Kurchatov would therefore merely suggest new directions
for research.37 It was perhaps not coincidental, then, that at the beginning of
1944 Lange was suddenly transferred to Sverdlovsk and Kikoin turned his
attention toward gaseous diffusion—a move that probably saved the enrich-
ment program considerable delay.38 Both Holloway and Alexei Kojevnikov
speculate that this might have been a mistake, because the centrifuge ulti-
mately proved more efficient.39 However, Lange’s horizontal centrifuge
would never have worked, and he showed no sign of abandoning his hori-
zontal predilections despite publications by Beams some six years earlier
showing that it was vertical machines that were needed.40 Like Beams, Lange
was married to his own ideas.41 It would take a new research group with a
fresh outlook before the centrifuge would become a practical machine.
The Means of Production
In the spring of 1945 the Red Army took Berlin and a special squad was
tasked with ferreting out German physicists and recruiting them into the
Soviet nuclear program. For those that agreed to help, laboratories were
established on the outskirts of Sukhumi, a small seaside resort town in
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 281
TECHNOLOGY AND CULTURE
APRIL
2012
VOL. 53
282
42. Specifically, two major laboratories—one under the leadership of Gustav Hertz
in Agudzheri and the other under Manfred von Ardenne in Sinop—were set up in what
are usually referred to as sanatoriums. Zippe, who was part of the research team, refers
to the buildings as Intourist hotels. For more on these activities, see Pavel V. Oleynikov,
“German Scientists in the Soviet Atomic Project”; Holloway, Stalin and the Bomb, 190;
and Gernot Zippe, Rasende Ofenrohre in stürmischen Zeiten, 73.
43. The accounts of Steenbeck’s transfer to the program report him as arriving ex-
tremely malnourished and weak, and it is tempting to read this as an indication of how
the Soviet’s treated their prisoners.According to Zippe,Steenbeck once confessed to him
that he tried to kill himself with what he thought was a cyanide pill upon being captured
by Soviet troops. Zippe speculates that this might have been the cause of Steenbeck’s per-
sistent stomach ulcers. See ibid., 75, and 137–38 (on Steenbeck’s personality); see also
Heinz Barwich and Elfi Barwich, Das Rote Atom, 38, 137–38.
44. For technical histories of the progress made while at the institute, see CIA,
“Development of Ultracentrifuge for Separation of Uranium Isotopes in the Soviet
Union” and “The Problem of Uranium Isotope Separation by Means of Ultracentrifuge
in the USSR”; and Zippe, “The Development of Short Bowl Ultracentrifuges” (reports
ORO-202, ORO-216, and ORO-315) and Rasende Ofenrohre in stürmischen Zeiten.
45. The influence of Beams’s early publications is noted in ibid., 85.
46. Ibid., 42. Holloway erroneously reports the name as “Konrad” Zippe in Stalin
and the Bomb, 191.
47. Sydney Evershed, “A Frictionless Motor Meter,” 743.
48. At that time, the formal theory of rotating shafts was only just being developed.
In the West this was being done by engineers like R. E. D.Bishop in the United Kingdom,
Georgia. Research staff members were supplied out of the scientists and
technicians imprisoned in the Soviet internment camps.42 One of these re-
searchers was Max Steenbeck—the former head of Siemens-Reiniger-
Werkes—an arrogant man who, in his words,“could afford it.”43 In Novem-
ber 1946, after a year of working on a dead-end project, Steenbeck asked if
he could work on the centrifuge. His wardens agreed and work began al-
most immediately, although Steenbeck did not receive his formal commis-
sion until February 1947.44 Unlike Lange, Steenbeck started with what
Beams had published and spent a year solving the problems inherent in the
Beams design through the thoughtful application of established solu-
tions.45 This he did with the help of his chief experimentalist, Gernot
Zippe, an Austrian physicist who had become a flight instructor in the Luft-
waffe shortly after completing doctoral work related to photoelectric ef-
fects.46 This team introduced three important features. The first was a
“point”bearing that allowed the centrifuge rotor to spin on the tip of a nee-
dle (like a toy top) with almost no friction. The idea was an adaptation of
the jewel bearing used in watchmaking and in Sydney Evershed’s electric-
ity meter developed in 1900.47 Steenbeck’s second major insight was the ap-
plication of loose bearings and weak damping, which allowed the cen-
trifuge to adjust itself so that it spun quietly on its center-of-mass axis
without vibration instead of trying to force the axis of rotation, as Beams’s
rigid bearings had done. This was an implementation of Carl Gustaf de
Laval’s principle of self-balancing, used for steam turbines since 1889.48
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 282
KEMPK|KGas Centrifuges and Nuclear Weapons
283
J. P. Den Hartog in the United States (who consulted for the U.S. centrifuge program),
and S. P. Timoshenko, a Russian émigré living in the United States. However, de Laval’s
principle, combined with Jeffcott’s 1919 rotor theory, was sufficient. See Carl Gustaf de
Laval, “Rotating Shaft,” U.S. Patent no. 431,750, issued 8 July 1890; and H. H. Jeffcott,
“The Lateral Vibration of Loaded Shafts,” 304. The experiments performed by Zippe for
Steenbeck are documented in Zippe, Rasende Ofenrohre in stürmischen Zeiten, 108ff.
Zippe believed that Soviet patience with these experiments was due, in part, to the fact
that the experimental work was advancing the state of basic knowledge about these phe-
nomena and was thus useful to Soviet science more generally (ibid., 111).
49. Beams’s centrifuges were driven directly by air jets against a turbine. Steenbeck’s
early designs used an electric motor and transmission. This idea for an indirect drive was
the result of an induction-welding system built by one of the POW Germans working
with Steenbeck, E. Steudel, who had been with Allgemeine Elektricitäts Gesellschaft
(AEG) prior to being captured (ibid., 104).
50. The dates of specific developments are reported in the 1957 interviews of Zippe
in CIA reports “Development of Ultracentrifuge for Separation of Uranium Isotopes in
the Soviet Union”and “The Problem of Uranium Isotope Separation by Means of Ul-tra-
centrifuge in the USSR,” as well as in later documents by Zippe, most comprehensively
in Rasende Ofenrohre in stürmischen Zeiten, 135, 144–46, 149, 151. The CIA reports
doubted Zippe’s accuracy on dates, but the purported dates appear to be consistent with
other historical information.
51. Holloway, Stalin and the Bomb, 191; I. N. Golovin, N. N. Ponomarev-Stepnoi,
and L. L. Sokolovskii, “On the 275th Anniversary of the Russian Academy of Sciences.”
52. In his original CIA interview Zippe claimed that Steenbeck had interactions with
Klaus Thiessen, who had developed a diffusion barrier at Sukhumi, and that Steenbeck
had learned from Thiessen that the barrier would not work at high enrichment levels.
Zippe suggested that Steenbeck thus wrote to Beria during the summer without any firm
knowledge that the Soviets had, in fact, used Thiessen’s barrier or that they were experi-
encing any difficulties. (See CIA report,“The Problem of Uranium Isotope Separation by
Means of Ultracentrifuge in the USSR,” 26.) The only supporting evidence for this claim
is that Steenbeck’s proposal was reportedly for a topping plant to enrich with an efficacy
of 50–90 percent, whereas Holloway, in Stalin and the Bomb, 191, reports that the diffu-
sion plant had only managed to reach 40 percent. These discrepancies, however, might
not be significant.
The third feature was to drive the rotation using electromagnetic fields, just
as the armature of an electric motor drives its internal rotating shaft. This
removed the need for a mechanical coupling and the hole and seal that
penetrated the vacuum casing in the Beams design.49 Together, these
changes solved essentially all the mechanical problems that had plagued
Manhattan Project centrifuges.50
In the summer of 1949 Soviet scientists discovered they also could not
exceed about 40 percent enrichment with their gaseous-diffusion plant—
just as the United States had discovered with its plant in 1943.51 Steenbeck
somehow learned of the problem and saw an opportunity.52 He wrote to
Lavrentiy Beria, People’s Commissar for Internal Affairs and overseer of the
Soviet nuclear program, with a proposal to build a topping plant that
would finish the job by enriching the uranium from 50 to 90 percent by
using centrifuges—exactly as Murphree had proposed during the Manhat-
tan Project. Steenbeck also wrote a second letter to Lieutenant General
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 283
TECHNOLOGY AND CULTURE
APRIL
2012
VOL. 53
284
53. Zippe, Rasende Ofenrohre in stürmischen Zeiten, 81, 136; CIA report, “The Prob-
lem of Uranium Isotope Separation by Means of Ultracentrifuge in the USSR,” 26–27.
54. The meeting reportedly involved about twenty officials, and it was evident that
the decision to build such a plant had already been taken. Steenbeck was invited only to
have his demands heard. See ibid., 27; Zippe, Rasende Ofenrohre in stürmischen Zeiten, 81.
55. Holloway, Stalin and the Bomb, 192.
56. CIA report, “The Problem of Uranium Isotope Separation by Means of Ultra-
centrifuge in the USSR,” 27, 32; Zippe, Rasende Ofenrohre in stürmischen Zeiten, 88.
57. CIA report, “The Problem of Uranium Isotope Separation by Means of Ultra-
centrifuge in the USSR,” 28.
58. Ibid., 30. On the difficulty caused by long centrifuges in Pakistan and Iran, see
R. Scott Kemp, “Nonproliferation Strategy in the Centrifuge Age.”
59. At the time, these bellows were called “sylphons.”
Avraamiy Pavlovich Zavenyagin, Beria’s deputy and longtime patron of the
German centrifuge work who had visited Steenbeck’s laboratory many
times. In this letter Steenbeck threatened that if the Soviet state did not now
make a commitment to release his family from captivity he would cease
work on the centrifuge altogether.53 At first the Soviets did not take this
coercion lightly and attempted to separate Steenbeck from the rest of the
research group, but eventually a meeting with Beria was arranged, proba-
bly in late July or August 1950, and an agreement reached.54
Holloway reports that in December 1950 the Technical Council criticized
Steenbeck’s proposal, thus putting an end to the topping-plant idea, but, in
fact, work on the plant continued, mainly focusing on developing a higher-
performing centrifuge that was also suitable for mass production.55 While
Steenbeck’s original group had solved the basic design problems of the cen-
trifuge, his prototype was still relatively expensive to build and operate
because it relied on external vacuum pumps to maintain the vacuum around
the centrifuge rotor, and compressors to pump uranium gas from one cen-
trifuge to the next. At an industrial scale these pumps would consume enor-
mous amounts of electricity, as they did in the gaseous-diffusion plants.
Steenbeck proposed—just as Beams had done in the U.S. program—to build
a very long centrifuge, about five meters in length.56 Longer machines are
capable of more separation, and this reduces the total number of machines
and their corresponding vacuum pumps and interstage compressors re-
quired, resulting in a more affordable plant. Steenbeck’s research staff was
expanded from twenty-five to sixty-five members, nearly all of them Rus-
sians, and new laboratory equipment and ample funding were provided.57
Steenbeck’s longer centrifuges, however, proved “incomparably more
difficult” to make—a property that would later cause significant strife for
Pakistan and Iran as well.58 They were built from interconnected tubes
joined by flexible “bellows” to compensate for the vibrations inherent in
longer tubes.59 The joints had to be tested and fine-tuned by hand, which
would have required either teams of talented tube-tuners or manufacturing
high-precision machine tools on a mass-production basis. However, the five-
meter centrifuge program never reached that point. Work on the new chal-
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 284
KEMPK|KGas Centrifuges and Nuclear Weapons
285
60. CIA report, “The Problem of Uranium Isotope Separation by Means of Ultra-
centrifuge in the USSR,” 29, 32; “A Modest Suggestion for a Review of the Bidding,”
129.
61.“Agreement between the Director of the Institute Kotschlawaschwili [sic] and the
Institute Laboratory Chief Dr. Gernot Zippe, 29 November 1950,” as reprinted in Zippe,
Rasende Ofenrohre in stürmischen Zeiten, 144–46. Kochlavashvili was Beria’s personal
representative (see Amy W. Knight, Beria, 138–39).
62. CIA report, “The Problem of Uranium Isotope Separation by Means of Ultra-
centrifuge in the USSR,” 27n18.
63. Also called OKB-133 for “Opytno-Konstruktorskoe Byuro 133.” Zippe, Rasende
Ofenrohre in stürmischen Zeiten, 163; CIA report, “The Problem of Uranium Isotope
Separation by Means of Ultracentrifuge in the USSR,” 36.
64. W. B. Gregory,“ Tests of Centrifugal Pumps”; R. G. Folsom,Review of the Pitot Tube.
65. The action of molecular pumps was first characterized by Gaede in 1913, and the
precise variation used by the Soviets was described by M. Holweck in 1923; see Holweck,
“Pompe moleculaire helicoidale.” The use of a Holweck molecular pump must have been
somewhat obvious, since Beams had independently suggested its use several years earlier
lenges continued at Sukhumi for about a year and a three-meter prototype
was being tested when, in December 1950, Soviet physicist Voskoboynik
solved the problem that had been plaguing the gaseous-diffusion plant,
thereby rendering the centrifuge topping plant superfluous.60 The program
nonetheless continued to receive support. Just a few weeks earlier, on 29 No-
vember 1950 Zippe and his staff technician Rudolf Scheffel, who were prob-
ably the only remaining Germans in the program, were asked to sign agree-
ments with the institute’s director, Ministry of Internal Affairs General
Aleksandr Kochlavashvili, to stay on and help the Soviets commercialize the
centrifuge in exchange for an earlier release home and handsome pay, which
they did.61 According to Zippe, Steenbeck had signed a similar agreement a
year earlier.62 The Soviet objectives were less limited than those of the Man-
hattan Project; they were focusing on the emerging arms race and the future
of commercial nuclear power, both of which required enrichment on a vast
scale. For the Soviets, the centrifuge still held considerable promise.
Work on the three-meter centrifuge continued with endurance tests
and efficiency improvements until 15 September 1952, when the entire
operation was moved to the Kirov Experimental Design Bureau (OKB-133)
in Leningrad, where compressors for the gaseous-diffusion plant were
being made.63 The centrifuge was being incorporated into the principal
programs of Soviet nuclear production.
In Leningrad, two Soviet engineers, Kharashavtsev and Nagorni, finally
solved the problem of expensive, energy-intensive pumps and compressors
by first adding a pitot tube to extract gas from the centrifuge. The tube har-
nesses the rotational momentum of the gas to pump it from one machine
to the next. This idea, which had been used for centrifugal pumps since
1901, eliminated the need for compressors.64 Their second addition was the
inclusion of a spiral-grooved sleeve around the rotor so that the outside of
the spinning centrifuge tube acted as a self-evacuating vacuum pump—an
implementation of the Holweck molecular pump invented in 1922.65 The
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 285
TECHNOLOGY AND CULTURE
APRIL
2012
VOL. 53
286
(in July 1952); see letters, Beams to T. H. Johnson, “Notes on KLI-1460 by Dr. G. F. Mills
on Isotope Separation,” and Beams to Johnson, 8 May 1952.
66. According to Zippe, the design was capable of one kilogram of 96-percent
enriched uranium metal per day using twenty kilometers of rotor tubing broken into
100-centimeter centrifuges, or 20,000 machines. This equates to 3.8 kg-SWU/year/ma-
chine. (See CIA report, “The Problem of Uranium Isotope Separation by Means of Ul-
tracentrifuge in the USSR,” 49.) However, it seems likely that this estimate assumed a
100-percent efficiency. Such a machine should normally have only about 0.8 kg-
SWU/year—almost one-fifth of Zippe’s estimate, but still adequate for a small-scale
weapons program. An SWU is the separative work unit—the amount of separation done
by an enrichment process. For performance-estimation methodologies, see Kemp,“Gas
Centrifuge Theory and Development.”
67. Kamenev headed a Soviet group that had been operating in parallel to the German
group since the early days of Sukhumi. The Germans were aware of Kamenev, as he and
his group regularly visited Sukhumi to learn of German advances, but the flow of infor-
mation was entirely one way. Kamenev’s group appears to have been originally located
elsewhere, perhaps in Moscow and not at the Kirov Experimental Design Bureau (OKB-
133). According to N. M. Sinev, who headed OKB-133 at the time of Steenbeck’s arrival,
Soviet centrifuge work (presumably at OKB-133) prior to the end of 1954 had been
headed by academician Boris P. Konstantinov and not Kikoin, who had been in charge of
enrichment under Kurchatov. See Oleg Bukharin, “Russia’s Gaseous Centrifuge Technol-
ogy and Uranium Enrichment Complex”; A. Plotkina, “The Development and Improve-
ment of the Centrifuge Method to Separate Uranium Isotopes in Russia”; Sinev, Enriched
Uranium for Atomic Weapons and Nuclear Power; and V.V. Shidlovsky and G. S. Soloviov,
“History and Status of Industrial Isotope Separation in Russian Federation.”
68. Shidlovsky and Soloviov, “History and Status of Industrial Isotope Separation in
Russian Federation.”
Leningrad group also added a magnetic bearing that further reduced fric-
tion, increased the machine’s lifetime, and reduced electricity consump-
tion. Again, the insightful application of existing solutions removed the re-
maining barriers to large-scale use; these advances eliminated the need to
pursue the difficult-to-make long machines advocated by Steenbeck. It was
now possible to build large numbers of simple, short machines on an eco-
nomically acceptable basis, which the Soviets did. The three Germans were
gradually eased out of the project by September 1954, but stayed just long
enough to witness the success of the first short machines in early testing.66
The “secret” of inexpensive, pragmatic centrifuges was now known to these
three non-Soviets, and what they would do with their knowledge would
dramatically shape the future of nuclear proliferation.
A variant of this machine was deployed on a pilot scale a few years later.
On 1 April 1954 a new group was established at Leningrad under the lead-
ership of Evgeni Kamenev to develop a commercial variant of the cen-
trifuge.67 The decision by the Council of Ministers to move forward with
the construction of a pilot plant was taken on 10 October 1955, and the
plant went into operation on 2 November 1957, reaching its full capacity of
2,435 centrifuges on 15 January 1958.68 The size was probably sufficient to
produce enough highly enriched uranium (HEU) for about one implo-
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 286
KEMPK|KGas Centrifuges and Nuclear Weapons
287
69. This is based on the performance of the SSZ-100 centrifuge, which the author esti-
mates to have had a separative power of about 0.77 kg-SWU/year, putting the pilot plant
at 1,500 SWU/year (accounting for cascade losses). For comparison,approximately 3,000–
8,000 SWU are needed for one nuclear weapon, depending on the design. The SSZ-100
was the second and last prototype machine of the Steenbeck-Kamenev type built by Zippe
in the Soviet Union, and it is assumed to be very similar to the machine used in the Soviet
pilot plant.It is possible that the design was changed slightly, but the timeline suggests that
such changes were likely to have been minor.For the methodology used in estimating the
performances of centrifuges, see Kemp, “Gas Centrifuge Theory and Development.”
70. J. Kistemaker, “Hoe een nieuwe industrie onstond (deel 2)”; N. L. Franklin,
“Looking Back to 1959”; W. E. Groth, K. Beyerle, E. Nann, and K. H. Welge, “Enrichment
of Uranium Isotopes by the Gas-Centrifuge Method.”
71. Franklin, “Looking Back to 1959.”
sion-type nuclear weapon every two years.69 This centrifuge plant was only
the first. The Soviet Union began to build more of them on increasingly
larger scales. By the end of 2008 Russia had expanded its centrifuge capac-
ity by a factor of 17,000 relative to the original 1957 pilot plant. Almost all
of this capacity was built during the cold war and gradually replaced the
country’s aging gaseous-diffusion plants—an indication of the tremendous
success of the Steenbeck–Zippe–Kamenev centrifuge.
The flawed centrifuge was made viable by the application of engineering
solutions that were mostly invented around the turn of the twentieth cen-
tury and all of which predated the Manhattan Project—evidence that the
latter’s centrifuge program was frustrated not by the limitations of manu-
facturing or the technology of the day, but rather by a preliminary design
that was never developed to its fullest possible extent.Whether this was due
to time constraints or a lack of insight cannot be stated, but the S-1 Uran-
ium Committee had pursued multiple enrichment technologies in parallel
to maximize the chances that at least one would be successful within the
Manhattan Project’s limited and time-sensitive objectives. What prevailed
was gaseous diffusion.
Revival in the West
Although there was little U.S.Atomic Energy Commission (AEC) inter-
est in the centrifuge after the Manhattan Project,other groups in the United
Kingdom, Germany, and the Netherlands continued to work on the Beams-
type centrifuge after the war and in parallel with Steenbeck’s work in the
Soviet Union.70 The United States, with its large gaseous-diffusion plants,
initially did not participate in such work; it had established itself as the pre-
dominant supplier of enriched uranium in the West and therefore had lit-
tle motivation to pursue alternative technologies. European researchers,
however, were looking for an affordable way to free their countries from the
United States’supplier monopoly, and to them centrifuges still appeared to
be the most promising small-scale technology.71
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 287
TECHNOLOGY AND CULTURE
APRIL
2012
VOL. 53
288
72. Kratzer interview.
73. AEC, “Gas Centrifuge Appraisal Report,” vol. 2. Kolstad to Kuhlthau, 27
December 1954. Redacted in 1985. Later documents suggest that the redacted item might
have been to explore other fissile-isotope separation—namely, purifying uranium-233
by removing uranium-232, as well as separating plutonium isotopes.
74. AEC, “Gas Centr ifuge Appraisal Report,” vol. 2.
75. CIA report, “The Problem of Uranium Isotope Separation by Means of Ultra-
centrifuge in the USSR.” The Germans working on isotope separation were part of an
agreement the Soviet Union made with its wartime allies that it be allowed to retain
within the country 1 million German specialists for ten years. The fact and dates of
Zippe’s and Steenbeck’s ultimate releases were according to this agreement.
By the early 1950s two West German groups—Konrad Beyerle’s in
Göttingen and Wilhelm Groth’s at the University of Bonn—had made con-
siderable progress with the rigid-bearing Beams-type design. To U.S. poli-
cymakers these devices appeared to pose more of a long-term commercial
rather than a proliferation threat; the West German devices were im-
mensely complicated and expensive and therefore plutonium still appeared
to be the proliferation route of choice. George Kolstad, chief of the AEC’s
physics and mathematics branch, began to advocate for restarting cen-
trifuge work on the basis that the AEC needed to understand the commer-
cial potential of the machines.72 In September 1954 the AEC formed an ad-
hoc committee to manage centrifuge affairs, whereby it was decided that
work should be undertaken in four domains: 1) studies on spinning very
long centrifuge rotors at the University of Virginia; 2) the manufacture and
successful operation of a long, supercritical Beams-type machine, with
preliminary studies at the Walter Kidde Nuclear Laboratories under the
direction of Karl Cohen and Beams, with the constructing contractor to be
chosen at a later date; 3) theoretical studies, also at Walter Kidde under the
direction of Cohen; and 4) an unknown item that was redacted from the
source documents at the time they were declassified.73
Actual research programs were not constituted until 1955. A contractor
search was completed, but no work on building machines was done nor
were the preliminary design studies undertaken at Walter Kidde.74 The AEC
Division of Research felt that the details of the West German design would
soon be made available to the United States, so it could bide its time; it was
content with limiting studies to those at the University of Virginia, which
occurred under the direction of Robert Kuhlthau, a former student of
Beams. Like the West Germans, the Kuhlthau group continued with the
Beams-type design; no effort was made to redesign the centrifuge, and with
the apparent progress in West Germany there was now even less motivation
to rethink the machine’s design limitations.
The redesigned and much-simplified Soviet centrifuge did not arrive in
the West until Steenbeck, Zippe, and Scheffel were released from captivity on
26 July 1956.75 After eleven years in the Soviet Union, Steenbeck rejoined his
family in the German Democratic Republic, where he became a professor of
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 288
KEMPK|KGas Centrifuges and Nuclear Weapons
289
76. M. Steenbeck, Impulse und Wirkungen.
77. Ronald D. Edge and Charles P. Poole, “Oswald Francis ‘Mike’ Schuette,” 85;
Houston G. Wood, ongoing personal communications with author, 2008; Kuhlthau,
“Notes by A. R. Kuhlthau from Discussion with Zippe,” “Updated Comments by A. R.
Kuhlthau about His Early Relationship with Zippe,” and Kuhlthau, personal interview
with author.
78. Kolstad was chief of the Physics and Mathematics Branch of the AEC’s Division
of Research.
79. Kuhlthau, “Notes by A. R. Kuhlthau from Discussion with Zippe,” “Updated
Comments by A. R. Kuhlthau about His Early Relationship with Zippe,” and interview.
U.S. intelligence was witting of the Soviet’s centrifuge project since at least 1955, but
lacked detailed information; see CIA reports,“Atomic Energy Research Work at Institute
‘C’ Headed by Manfred von Ardenne” and “Isotope Separation at the Hertz Institute.”
80. CIA report, “The Problem of Uranium Isotope Separation by Means of Ultra-
centrifuge in the USSR.”
plasma physics at Friedrich Schiller University in Jena.76 Both Zippe and
Scheffel, with a more capitalistic drive, headed to the West. Within weeks of
his return Zippe crossed paths with Oswald Francis“Mike”Schuette, a young
American physics professor on a Fulbright scholarship at the Max Plank
Institute in Mainz and who was also working as an agent for the Office of
Naval Intelligence.77 Schuette learned of Zippe’s work in the Soviet Union
and wrote a report about it that piqued the interest of Kolstad at the AEC,
who arranged to have Zippe more thoroughly interrogated.78 Accompanied
by Schuette, Zippe agreed to come to the United States in early 1957, travel-
ing with a false passport and under the assumed name of “Dr. Schubert.”
Kolstad, Schuette, Kuhlthau, and a number of intelligence agents interviewed
Zippe in a day-long session at the Shoreham Hotel in Washington, D.C.79
The two reports of this interview describe Zippe as being“candid by nature”
and together comprise 119 pages, including the entire history of his time
spent in the Soviet Union, drawings of the various centrifuges built by the
German team, the names of the people involved, and even the floor plans of
the Kirov plant where he last worked.80 Most importantly, Zippe explained
in detail how to solve the problems of the Beams-type centrifuge.
At some point between the spring of 1957 and the spring of the follow-
ing year Kolstad decided he needed Zippe back to re-create the Soviet cen-
trifuge for the U.S. government. Zippe agreed in May 1958 and the AEC
therefore created Project 2400, which provided money to the University of
Virginia, which would, in turn, hire Zippe as a consultant. Zippe arrived in
Virginia in August. The project was administered by A. Robert Kulhthau,
who, unknown to Zippe, was also concurrently working on classified AEC
centrifuge work. As a foreign national Zippe was kept isolated from these
other U.S. centrifuge activities. He had at his disposal one machinist, whom
he brought with him from Germany; the university’s resident fluid-dynam-
ics theorist (for whom he had little use); William Dancy, a young mechanic
“who was good with his hands” and later ran the old Nier mass spectrom-
eter to analyze samples of enriched uranium; and Wilbur May, a former
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 289
TECHNOLOGY AND CULTURE
APRIL
2012
VOL. 53
290
81. Kuhlthau interview; quote from Zippe, The Development of Short Bowl Ultracen-
trifuges, 5.
82. The first test stands, used for balancing operational systems in atmospheric air,
as well as for runs in high vacuum, were completed and in use by October 1958.
Preparatory bearing- and rotor-lifetime tests began in November 1958. A spin test stand
for determining the bursting speed of materials and for testing new designs of rotor end-
caps and baffles went into operation in June 1959. Tests to determine the drag from the
scoops were made from February through mid-April 1959. Tests on different Holweck-
type molecular pumps with different gases (enabling an extrapolation to UF6) occurred
from mid-April through July 1959. Separation tests with UF6gas were delayed, because
of a delay in the delivery of UF6, but were conducted with Freon in the interim. The
actual separation of UF6commenced in September 1959 and continued until the termi-
nation of the program in May 1960.
83. Performance results are from Kuhlthau,“Check on Zippe Data.”
84. Two standing AEC contractors were invited to see Zippe’s work: General Electric
Laboratories on 5 July 1960, and Union Carbide on 6 July (R. A. Lowry, “Laboratory
Notebook 7,” 6 July 1960). The Dow Chemical Company had been approached for a pos-
sible collaboration by the German firm DEGUSSA, which had been developing the
Soviet-type centrifuge in Germany, and representatives from Dow were briefed on the
technology (Lowry, “Laboratory Notebook 7,” 7 April 1960). Also briefed were Nobel-
laureate Maria Goeppert-Mayer and J. Newgard, who as a result founded a private cen-
trifuge-enrichment company called Electro-Nucleonics to separate tungsten isotopes,
the AEC license for which was later denied, therefore it turned its attention to the devel-
opment of the biological centrifuge. Hans Kronberger, director of research and develop-
ment for the U.K.Atomic Energy Agency, was also briefed on Zippe’s centrifuge and con-
sequently the U.K. government was encouraged to give the centrifuge another try. See
Zippe, “Unclassified Spots on the History of the Modern Gas Centrifuge.”
army cook who assisted Zippe with experiments. Zippe wrote in his final
report: “A small machine shop has been installed in this laboratory build-
ing. This was suitable for about 90% of the mechanical work required for
the project. The remaining 10%, usually requiring heavier machinery, has
been done either in the Physics Department shop or in the main shop of
the Research Laboratories for the Engineering Sciences.” The resources
available to Zippe were therefore modest.81
Zippe’s project ran for nearly two years, fromAugust 1958 to June 1960,
and included several months of unproductive time as a result of the diffi-
culty in procuring uranium.82 The machine he built was smaller than his
Soviet one. It performed at 0.43 SWU/year, or about half of what one might
ideally use for a small weapons program.83 Zippe wrote a report detailing
the development of this machine, and experts at the AEC, as well as at sev-
eral private firms, marveled at the device’s simplicity.84
It comes as no surprise that Zippe was able to reproduce the machine
in such a short time. Although he had not worked on centrifuges for four
years, he had long experience with them and probably possessed all the
tacit knowledge required. But how long would it take an inexperienced
group of engineers to build a centrifuge when their only starting point was
a general knowledge of the design? The subsequent development program
run by the AEC provides one answer to this question.
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 290
KEMPK|KGas Centrifuges and Nuclear Weapons
291
85. He suggested that the funding for the short (subcritical) centrifuge program be six
times larger than the funding for the long (supercritical) centrifuges under the direction of
Kuhlthau at the University of Virginia; see letter, Karl Cohen to Beams, 1 February 1960.
86. Research Laboratories for the Engineering Sciences,University of Virginia,“Pro-
posal for the Extension of Contract AT-(40-1)-1779.”
87. This decision was made in April 1960; see AEC, “Gas Centrifuge Appraisal Re-
port,” vol. 2.
88. The contract was initially awarded to General Electric Laboratories in Schenec-
tady, New York, where Karl Cohen was employed and already exploring centrifuge
options. It was awarded at some point after July 1960, then terminated within a few
months for unknown reasons. The Union Carbide Nuclear Company (UCNC) had ear-
lier secured, at some point before May 1960, a contract to explore centrifuges with the
CIA. The cancellation of GE’s contract and its reassignment to UCNC may have been in-
fluenced by a desire to consolidate efforts.
89. The program had four declared objectives: 1) to ascertain the potential for fur-
ther improvements of the centrifuge process; 2) to improve the accuracy of economic
projections; 3) to estimate the potential of small nations to produce enriched uranium
for weapons purposes; and 4) to evaluate the ability of gas centrifuges to separate other
isotopes, including the removal of 232U from 233U and the isotopes of plutonium. See
UCNC, “Proposal for the Development of the Gas Centrifuge Process of Isotope
Separation.”
The U.S. Centrifuge Program
The United States’effort to build a reliable, high-performing centrifuge
based on the Soviet design did not commence in earnest until the autumn
of 1960. During the previous winter the AEC had asked Karl Cohen to
review Zippe’s work at the University of Virginia. Cohen was deeply im-
pressed by the simplicity of Zippe’s device and recommended in February
1960 that the country establish a sizable program based on its design.85
This recommendation ran counter to the preferences of researchers at the
University of Virginia, who were still devotees of the Beams design.86
Therefore the AEC decided on a two-track approach: it would let the Vir-
ginia group continue studies on the Beams-type centrifuge, while concur-
rently awarding a new contract for the ongoing development of the Soviet-
type centrifuge.87
This contract was awarded to the Union Carbide Nuclear Company
(UCNC), the firm managing the Oak Ridge gaseous-diffusion plant.88
Strictly adhering to the AEC’s mandate, UCNC proposed to demonstrate a
centrifuge capable of spinning at 450 meters per second; and to operate an
experimental cascade of centrifuges in order to understand their prolifera-
tion aspects and the economics of centrifuge enrichment.89 It was not in-
tended that UCNC develop a commercial centrifuge design to replace the
gaseous-diffusion plant it managed at Oak Ridge.
Given that the United States had long studied centrifuges (although of
a different type) and also given the demonstration of them by Zippe at the
University of Virginia, it is important to ask in what ways UCNC’s program
benefited—as it turns out, less than might be expected. First of all, UCNC
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 291
TECHNOLOGY AND CULTURE
APRIL
2012
VOL. 53
292
did not have the benefit of Zippe himself, because he had left the country
the previous June and for security reasons was no longer able to act as con-
sultant to the U.S. program.90 His machinist Mr. Deutsch had also returned
to Germany. While other centrifuge researchers at the University of Vir-
ginia did have the opportunity to observe Zippe at work, they were other-
wise engaged in their own research activities. The monthly progress reports
from the Virginia group did not mention any other individuals working on
the Zippe design either before or after his departure; these reports were
mostly dismissive of Zippe’s ideas.91 The only person with hands-on expe-
rience working with Zippe remaining in the United States was Dancy, who
had assisted him in miscellaneous activities.92 After Zippe’s departure
Dancy continued to work at the University of Virginia in building motors
and instruments for the Beams-type centrifuge project. Beyond occasional
consultations with Dancy, most of the knowledge about the Soviet design
was conveyed by the very detailed technical reports Zippe wrote during the
final days of his project.93
UCNC was potentially advantaged because it could share resources
with the Virginia program; however, the value of this cooperation would
have been limited. Recall that the Beams-type centrifuge did not have a
scoop system and used a different type of bearing; consequently,most of its
critical parts bore no resemblance to Zippe’s design.94 Among the primary
commonalities would have been the molecular pumps and some external
gas-handling hardware (for example, pipes and valves), although the design
of the pump was neither sensitive nor well understood by the Virginia
group.95 As to theory, researchers at the University of Virginia had been
working on centrifuge fluid dynamics and rotor dynamics, and they would
have been able to give UCNC’s engineers crash courses on the subjects.
However, it turns out that their understanding of these phenomena was
90. Lowry,“Laboratory Notebook 7,” 25 March 1960. Zippe reportedly went back to
Germany because of his fiancée and to continue his earlier contract employment with
DEGUSSA, which was also developing a centrifuge at the time.
91. Research Laboratories for the Engineering Sciences, “Gas Centrifuge Progress
Summary, 1 June 1960 through 31 August 1960” and “Quarterly Progress Report for the
Period 1 June 1960 to 1 September 1960.”
92. Kuhlthau interview.
93. Zippe, Beams, and Kuhlthau, “The Development of Short Bowl Ultracentri-
fuges”; Zippe, “The Development of Short Bowl Ultracentrifuges” (reports ORO-202,
ORO-216, and ORO-315).
94. The Beams-type centrifuge had been improved by moving to hydrosphere-type
bearings. Countercurrent pumping was achieved through an external pumping and spe-
cially designed end-caps, and later with rotating brake-discs, that approximated the
angular-momentum loss of a scoop system.
95. Lowry, “UCNC Centrifuge Work,” dated 8 March 1961, in “Laboratory Note-
book 15.” A later publication of the University of Virginia group reveals that its molecu-
lar-pump model was flawed; see E. N. Sickafus, R. B. Nelson, and Lowry,“The Holweck
Type Molecular Pump.”
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 292
KEMPK|KGas Centrifuges and Nuclear Weapons
293
96. Six-inch-diameter tubes were produced by UCNC prior to 30 April 1961; see E. C.
Evans and E. F. Babelay, “Gas Centrifuge Development, Progress Report, November 1,
1960 through April 30, 1961.” The machine shop at theY-12 facility began manufacturing
tubes for the program in September 1961 (letter, Lowry to Babelay, 25 September 1961).
97. S. W. Palmer, “Notes on the Roles of Early Members of the UCCND Gas Cen-
trifuge Development Group.”
98. UCNC, “Gas Centrifuge Development of Process Development [sic], Quarterly
Report for Third Fiscal Quarter, Jan 1, 1961 – March 31, 1961.”
99. Evans and Babelay, “Gas Centrifuge Development, Progress Report, November 1,
1960 through April 30, 1961.”
100. UCNC,“Gas Centrifuge Development of Process Development [sic], Quarterly
Report for Third Fiscal Quarter, Jan 1, 1961 – March 31, 1961”; Evans and Babelay,“Gas
Centrifuge Development, Progress Report, November 1, 1960 through April 30, 1961.”
101. UCNC,“Gas Centrifuge Development of Process Development [sic], Quarterly
Report for Third Fiscal Quarter, Jan 1, 1961 – March 31, 1961.”
either flawed or underdeveloped and thus could have only introduced
delays in the process. The most useful cooperation appears to have been in
the procurement of rotor tubes. The Virginia program had been negotiat-
ing with a handful of industrial metal firms for several years regarding such
tubes, and thus UCNC was able to tap into this supplier network to obtain
tubes necessary for its first batch of centrifuges, although it began produc-
ing its own tubes within six months.96 In sum, UCNC’s benefit of having
access to earlier and ongoing research on Beams-type centrifuges would
have been marginal at best.
UCNC’s development work on centrifuges began at Oak Ridge on 1
November 1960, initially being carried out by a group of four.97 By the end
of March 1961, after the first five months of the program, this small group
accomplished nearly all of the design and testing needed to build a func-
tioning machine. Material studies on the metallurgy, forming, machining,
corrosion resistance, and creep of the high-strength aluminum used for
rotor bodies, and also of the tool steel used in the needle bearing, were
completed.98 Sample rotors and bearing configurations were then built and
mechanically tested and used to validate the accuracy of diagnostic tools, as
well as to measure the motor drive’s characteristics and the performance of
the molecular pump.99 A range of different needle-bearing and seat mate-
rials was explored, and two different lower-suspension systems employing
different damping mechanisms and two different end-cap designs were
then evaluated to determine which provided the optimal mechanical per-
formances.100 The group also learned how to make consistently accurate
(namely, within plus or minus 0.15 percent) measurements of the isotopic
content of uranium samples using a mass spectrometer,a necessary step for
fine-tuning the separation performances of centrifuges.101 With these pro-
cedures in place, the group then embarked on a series of over 300 separa-
tion experiments using stand-alone, three-inch-diameter centrifuges to
determine the optimal uranium feed rates of each of twelve different scoop
and baffle configurations, and the best performing of these were selected
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 293
TECHNOLOGY AND CULTURE
APRIL
2012
VOL. 53
294
102. Ibid.
103. These achievements were probably expedited by the fact that UCNC could
draw on staff members with experience in operating a mass spectrometer from the
gaseous-diffusion plant, and metallurgists familiar with handling UF6and UF6metal re-
actions. Although the initial four engineers had no experience with centrifuges or related
technologies, they were familiar with gas-diffusion-related technologies, rudimentary
separation theory, and handling uranium.
104. For titanium: UCNC, “Gas Centrifuge Development of Process Development
[sic], Quarterly Report for Third Fiscal Quarter, Jan 1, 1961 – March 31,1961”; for exotic
aluminum alloys: Lowry, notes dated 9 and 10 November 1960, in “Laboratory Note-
book 12”; for fiberglass: Evans and Babelay, “Gas Centrifuge Development, Progress
Report, November 1, 1960 through April 30, 1961”;and for beryllium: Lowry, note dated
21 December 1960, in “Laboratory Notebook 12.”
105. This decision was made on about 15 December 1960; see Lowry, notes dated 20
and 21 February 1961, in “Laboratory Notebook 12.” The performance of a centrifuge
increases linearly with length, but is independent of pure changes in the diameter; how-
for the prototype.102 Finally, the group drew up plans for a thirty-five-ma-
chine centrifuge cascade.103
At this point, the production of many centrifuges was needed to test
their performances and therefore new staff members were added, thus ex-
panding the program from four to fifteen people, with most starting dur-
ing the second and third quarters of the program. A close review of these
additional staff members’ academic resumes indicates that they had not
studied at the elite institutions usually associated with the early days of the
Manhattan Project, but were, rather, competently trained or untrained en-
gineers or technicians; only two of the new eleven had master’s degrees in
those fields, and one had no college education at all.All of this suggests how
readily such a program might be emulated today even in places where per-
sonnel resources are limited, as they are in some developing countries.
As UCNC’s centrifuge program grew in size, its staff members were
spread over a greater number of research tasks, among which were several
related to the commercialization of the device. For example, the quest for
high-speed enrichment, which was deemed necessary for economic com-
petitiveness, compelled researchers to perform additional metallurgical
studies on materials capable of operating under greater stresses, including
beta titanium, exotic aluminum alloys, fiberglass, and beryllium.104 The
program also spent additional time in designing a dry lower bearing, be-
cause the lubricating oils available at the time were not chemically com-
patible with uranium hexafluoride (UF6) and would eventually corrode.
Chemically problematic lubricants would have increased maintenance
costs in the long term—another concern for commercial centrifuges,
though not necessarily one for a basic, proliferation-capable centrifuge.
Today, UF6-resistant lubricants are readily available.
By the end of its second quarter the program had also begun work on
a second-generation centrifuge with a larger, five-inch-diameter rotor that
would nearly double its performance.105 Two of these were being built at
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 294
KEMPK|KGas Centrifuges and Nuclear Weapons
295
ever, the maximum allowable length is linearly dependent on the diameter, so increasing
the diameter permits a longer machine and the performance increases accordingly.
106. UCNC,“Gas Centrifuge Development of Process Development [sic], Quarterly
Report for Third Fiscal Quarter, Jan 1, 1961 – March 31, 1961”; Evans and Babelay,“Gas
Centrifuge Development, Progress Report, November 1, 1960 through April 30, 1961.”
107. Palmer, “Notes on the Roles of Early Members of the UCCND Gas Centrifuge
Development Group.”
108. Machines had been tested to 370 meters per second, and vibrations appeared to
be around 320 meters per second; see J.H. Dickerson and Lowry,“Notes on the Mechan-
ical Performance of 5-inch Diameter Gas Centrifuge Rotors.”
109. UCNC,“Gas Centrifuge Development of Process Development [sic], Quarterly
Report for First Fiscal Quarter, July 1, 1961 – September 30, 1961.”
110. Palmer, “Notes on the Roles of Early Members of the UCCND Gas Centrifuge
Development Group.”
111. Evans and Babelay, “Gas Centrifuge Development, Progress Report, May 1,
1961 through November 30, 1961.”
112. UCNC,“Gas Centrifuge Development of Process Development [sic], Quarterly
Report for Second Fiscal Quarter, October 1, 1961 – December 31, 1961.”
the end of the sixth month, and it was thought that they might be used for
the planned test cascade instead of the three-inch-diameter machines.106
Four researchers were assigned to the machine assembly and design effort,
and four others to the cascade effort.107 The cascade’s design was com-
pleted, and 95 percent of the required equipment had been delivered by 30
June 1961. It was during this period that the program experienced its first
major problem. As mentioned above, in the hope of improving the cen-
trifuge’s performance the program had moved from a three-inch-diameter
to a five-inch-diameter rotor. Although the latter centrifuge had operated
successfully during a preliminary test in May, additional five-inch-diame-
ter rotors produced for the cascade exhibited unpredictable vibrations near
operating speed.108 Consequently, it was decided to continue with the con-
struction of the cascade using the three-inch-diameter design instead.109
Progress on the cascade proceeded during the program’s fourth and
fifth quarters. Five engineers were assigned to manufacture components for
the three-inch-diameter centrifuge, which they did with the help of four
additional technicians. The cascade design team brought on two additional
engineers to handle the balance of the plant—namely, the vacuum systems,
instrumentation, and cascade controls. Two more technicians helped with
miscellaneous mechanical tasks.110
During early tests the three-inch-diameter centrifuges also began to
reveal vibration problems; of twenty-eight built, only twelve survived test-
ing.111 A small test cascade using the twelve surviving devices was placed in
operation on 30 November and was successful. Over the next thirty days,
twenty cascade tests were carried out and the cascade’s efficiency improved
from 65 to 80 percent.112
The vibration problem first experienced with the five-inch-diameter
machines and then later with the three-inch ones is of particular interest,
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 295
TECHNOLOGY AND CULTURE
APRIL
2012
VOL. 53
296
113. Namely, the building cascade infrastructure and running cascades tests, but also
further studies on motor, bearings, and end-caps, and planning for a third-generation,
twenty-four-inch-diameter centrifuge. See Evans and Babelay,“Gas Centrifuge Develop-
ment, Progress Report, May 1, 1961 through November 30, 1961”; and UCNC, “Gas
Centrifuge Development of Process Development [sic], Quarterly Report for Second Fis-
cal Quarter, October 1, 1961 – December 31, 1961.”
114. Lowry, “Laboratory Notebook 7.”
115. Dean A. Waters, “A Vibration Study of Subcritical Gas Centrifuges.” An even
more complete and accurate theory had been published two years earlier, but they
seemed unaware of its existence; see R. E. D. Bishop,“The Vibration of Rotating Shafts.”
116. UCNC,“Gas Centrifuge Development of Process Development [sic], Quarterly
Report for Third Fiscal Quarter, Jan 1, 1962 – March 31, 1962.”
because official program records suggest that it was the only significant
technical hurdle in the development of a reliable centrifuge. There was no
prior experience to turn to for guidance because the problem was unique
to the Soviet, and not to the Beams, design. In about August 1961 a staff
member named Dean Waters, age 25, was assigned to study the vibration
problem, the remainder of the staff being devoted primarily to cascade
tasks.113 Waters carefully cataloged the vibration modes of the three-inch-
diameter centrifuge over the period of a month, finding that many of the
resonances had not been predicted by the program’s simple theory of rotor
dynamics. He then enlisted the help of Joe Bodine of the University of Vir-
ginia to devise an improved theory of rotor dynamics.114 Waters’s report
was released on 3 October 1961.115 In January 1962, during the program’s
sixth quarter, the vibration problem of the three-inch-diameter rotors was
solved by tuning the damping of the bearing support system to the optimal
level predicted by the report’s improved theory of rotor dynamics.116
By the early spring of 1962,fifteen months after the program’s inception
and with a team never exceeding fifteen members, the UCNC program had:
1) mastered the mechanical difficulties of the centrifuge; 2) learned how to
manufacture all components in-house, including rotor tubes; 3) demon-
strated that it could repeatedly build reliable centrifuges; and 4) designed,
assembled, and operated a centrifuge cascade to an efficiency of 80 percent.
The role of individual preference proved to be very important in both
the U.S. and Soviet programs. In the United States, Beams’s preference for
rigid-bearing designs dominated U.S. centrifuge work for approximately
thirteen years, from 1937–44 and 1953–59, and after the war the Germans
followed suit until Zippe’s design changed their perspective. Similarly, the
Soviet Union’s enrichment program run by Kikoin was, from 1942 to 1944,
focused mainly on Lange’s horizontal centrifuge. It wasn’t until Kurchatov
later learned of the U.S. decision to build a full-scale gaseous-diffusion
plant that he shifted the Soviet focus to gaseous diffusion, albeit only tem-
porarily. Thus while spying might not have accelerated the overall pace of
the Soviet program by much (its timeline being instead dominated by the
plutonium program), it does appear to have kept the program from being
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 296
KEMPK|KGas Centrifuges and Nuclear Weapons
297
117. Zippe, Rasende Ofenrohre in stürmischen Zeiten.
118. Shidlovsky and Soloviov, “History and Status of Industrial Isotope Separation
in Russian Federation.”
119. The most comprehensive history of the early British program is given in Kemp,
“Nonproliferation Strategy in the Centrifuge Age.”
120. T. Trevor Edwards, personal interview with author.
121. Computed from the installed machine capacities given in International Atomic
Energy Agency,“Implementation of the NPT Safeguards Agreement and Relevant Pro-
visions of Security Council Resolutions 1737 (2006), 1747 (2007) and 1803 (2008) in the
Islamic Republic of Iran” and “Implementation of the NPT Safeguards Agreement and
Relevant Provisions of Security Council Resolutions 1737 (2006), 1747 (2007), 1803
(2008) and 1835 (2008) in the Islamic Republic of Iran.”
focused solely on Lange’s centrifuge. Similarly, during construction of the
Soviet’s topping plant Steenbeck championed the building of longer, five-
inch-diameter centrifuges, even heatedly arguing with his subordinate
Zippe over the wisdom of longer versus shorter centrifuge designs.117 Ul-
timately, it was Soviet engineers who realized the superiority of the shorter,
three-inch-diameter machines.
The UCNC program provides an excellent sense of the development
effort required to replicate the Soviet-type centrifuge, which started with
only Zippe’s reports. The research staff and effort involved were surprisingly
small: a maximum of fifteen people taking less than fifteen months. Signif-
icantly, there appears to be no meaningful transfer of tacit knowledge from
Zippe to the UCNC program. However, the UCNC program is not inform-
ative regarding the effort required to build a pilot plant itself, because it
never intended to take that step—but other countries did. The Soviet Union
finalized its 2,400-machine plant twenty-seven months after its program
was launched, although its staffing details are unknown and probably
large.118 Also instructive is the example of the British pilot plant, which was
built at short notice in an attempt to out-compete a parallel gaseous-diffu-
sion pilot program.119 According to a Trevor Edwards, an engineer who
worked on Great Britain’s first pilot centrifuge cascade, approximately 2,000
machines were assembled in about one year. Furthermore, because time was
of the essence and resources were scarce, the manufacturing operation was
improvised by hiring teams of unskilled workers to machine and assemble
parts in a production line; also,solutions to some tasks were jury-rigged (for
example, by using audio amplifiers as power supplies).120
Iran’s present centrifuge program has also been informative. Although
the country’s centrifuges are far more complicated to manufacture than the
short Soviet-type devices discussed above, Iran nonetheless manages to
produce about 1,500 annually.121 All these estimates are more or less con-
sistent, suggesting that the 2,000–5,000 machines needed for a small weap-
ons program might take an additional one to three years to build once a
centrifuge prototype was established.
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 297
TECHNOLOGY AND CULTURE
APRIL
2012
VOL. 53
298
122. For example, in the Franck Report (“Report of the Committee on Political and
Social Problems”).
123. Truman, “Statement by the President.”
Conclusions
There has been a long-standing belief among policymakers in the
United States that with the success of the Manhattan Project came a hege-
mony in the field of atomic weapons, especially with respect to uranium
enrichment. While the physicists involved in the project had warned that
“nuclear bombs cannot possibly remain a ‘secret weapon’ at the exclusive
disposal of this country for more than a few years,” the obviously vast scale
of the Manhattan Project seemed too enormous for other nations to take
on, especially when it came to uranium enrichment.122 As President
Truman said, it was not just “the achievement of scientific brains . . . hardly
less marvelous has been the capacity of industry to design, and of labor to
operate, the machines and methods to do things never done before.”123
On the surface, history would appear to have proved the politicians
right, but this article has drawn very different boundaries for the infrastruc-
ture needed for nuclear proliferation. The history of the Soviet centrifuge
program shows that established solutions were able to convert failed Man-
hattan Project centrifuges into simple, easy-to-make machines. The subse-
quent replication of them by UCNC provides conclusive evidence that the
scale involved can indeed be small: fifteen persons, with no prior experience
or specialized knowledge, taking less than fifteen months to achieve success.
Other countries’ program histories also suggest that mastery of the mass-
production process of centrifuges is not particularly difficult.
If such an uncomplicated, small-scale pathway has long been available,
then why did countries not follow it? This article recounts that the Man-
hattan Project’s management decided to abandon the centrifuge, because
there appeared to be too many remaining problems with it and, consider-
ing the time constraints, a decision had to be made—which, for better or
worse, favored gaseous diffusion. Similar problems existed in the early
Soviet program, but parallel research teams ultimately determined that
centrifuges were feasible. The centrifuge was, in principle, a viable technol-
ogy at the time of the Manhattan Project, but it was not pursued simply be-
cause of individual biases and organizational constraints. This raises a
question: What happened in the states that built nuclear weapons after
World War II, when such time pressures were relaxed?
This article has not explored the French, Israeli, Chinese, or Indian
nuclear programs, but the decisions of Kurchatov would appear to suggest
that, under intense internal pressure to perform, program managers can be
risk-averse and turn away from technologies that physics suggest are opti-
mal toward those for which there is at least a“proof of existence.” The Man-
hattan Project provided one such proof,although a highly inferior one, and
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 298
KEMPK|KGas Centrifuges and Nuclear Weapons
299
the first six proliferators that followed replicated its technology choices
instead of reevaluating the field of the possible. Had the United States been
successful in establishing the centrifuge model, and had it become known
that centrifuges were as viable as they were, there might have been consid-
erably more proliferation starting in 1945 and continuing up to the time at
which a formalized system of political control could be established. As his-
tory would have it, though, political control was not established until 1970
with the signing of the Treaty on the Non-Proliferation of Nuclear Weap-
ons. Such a treaty does not, however, stop countries from experimenting
with centrifuges as a backup plan. Today, at least twenty countries have
built or acquired centrifuge technology, and the history lesson drawn here
suggests that it is within the capability of nearly any state to do so. This
should serve to remind policymakers that they need to remain good stew-
ards of the political barriers to proliferation, for the technological ones are
weak at best.
Bibliography
Published and Unpublished Works Not from Archive Collections
Albright, David. Peddling Peril: How the Secret Nuclear Trade Arms Ameri-
ca’s Enemies. New York: Free Press, 2010.
Barwich, Heinz, and Elfi Barwich. Das Rote Atom. Klagenfurt, Austria: Ed-
uard Kaiser Verlag, 1967.
Beams,Jesse W.“High Rotational Speeds.” Journal of Applied Physics 8 (1937):
795–806.
———. “High Speed Centrifuging.” Reviews of Modern Physics 10 (1938):
245–63.
———. Early History of the Gas Centrifuge Work in the U.S.A. Charlottes-
ville: Department of Physics, University of Virginia, 1975.
———, A. C. Hagg, and E. V. Murphree. Developments in the Centrifuge
Separation Project. Oak Ridge, TN: U.S. Atomic Energy Commission,
1951.
———, and F. B. Haynes. “The Separation of Isotopes by Centrifuging.”
Physical Review 50 (1936): 491–492.
Beyerle, K., P. Harteck, W. Groth, and H. Jensen. Über Gaszentrifugen. Bei-
hefte zu Angewandte Chemie und Chemie-Ingenieur-Technik 59. Wein-
heim, West Germany: Verlag Chemie GMBH, 1950.
Bishop, R. E. D. “The Vibration of Rotating Shafts.” Journal of Mechanical
Engineering Science 1 (1959): 50–65.
Bukharin, Oleg. “Russia’s Gaseous Centrifuge Technology and Uranium
Enrichment Complex.” Working paper, Program on Science and Global
Security, Princeton University, January 2004.
Central Intelligence Agency. “Isotope Separation at the Hertz Institute.”
Report, 5 February 1955.
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 299
TECHNOLOGY AND CULTURE
APRIL
2012
VOL. 53
300
———. “Atomic Energy Research Work at Institute ‘C’ Headed by Manfred
von Ardenne.” Report, CIA-RDP80-00810A006600620008-6, 11 Octo-
ber 1955.
______. “Development of Ultracentrifuge for Separation of Uranium Iso-
topes in the Soviet Union.” Scientific and Technical Information Branch,
Report EG-1795, 18 September 1957.
———. “The Problem of Uranium Isotope Separation by Means of Ultra-
centrifuge in the USSR.” Scientific and Technical Information Branch,
Report EG-1802, 8 October 1957.
———. “Communist China’s Advanced Weapons Program.” Special Na-
tional Intelligence Estimate 13-2-63, 24 July 1964.
Collins, Catherine, and Douglas Frantz. Fallout: The True Story of the CIA’s
Secret War on Nuclear Trafficking. New York: Free Press, 2011.
Corera, Gordon. Shopping for Bombs: Nuclear Proliferation, Global Insecur-
ity, and the Rise and Fall of the A.Q. Khan Network. New York: Oxford
University Press, 2006.
Dangdai Zhongguo (China Today), ed. “Research on Nuclear Science and
Technology, Section 5: Research on Uranium Isotope Separation.” In
Modern China’s Nuclear Industry, translated by Joint Publication Re-
search Service, edited by Dangdai Zhongguo, 387–91. Rosslyn, VA: For-
eign Broadcast Information Service, 1988.
De Laval, Carl Gustaf Patrick. “Rotating Shaft.” U.S. Patent no. 431,750,
issued 8 July 1890.
Edge, Ronald D., and Charles P. Poole. “Oswald Francis ‘Mike’ Schuette.”
Physics Today 53 (2000): 85.
Edwards, T. Trevor. Personal interview with author, 10 March 2008, Vienna.
Evershed, Sydney. “A Frictionless Motor Meter.” Journal of the Institute of
Electrical Engineers 29 (1900): 743–94.
Folsom, R. G. Review of the Pitot Tube. Chicago: American Society of Me-
chanical Engineers, 1955.
Franck, James, Donald J. Hughes, J. J. Nickson, Eugene Rabinowitch,
Glenn T. Seaborg, J. C. Sterns, and Leo Szilard. “Report of the Commit-
tee on Political and Social Problems” (Franck Report). Manhattan
Project Metallurgical Laboratory, University of Chicago, 11 June 1945.
Franklin, N. L.“Looking Back to 1959.” The Nuclear Engineer 26 (1985): 1–
12.
Frantz, Douglas, and Catherine Collins. The Nuclear Jihadist: The True
Story of the Man Who Sold the World’s Most Dangerous Secrets . . . and
How We Could Have Stopped Him. New York: Twelve, 2007.
Furry, W. H., R. Clark Jones, and L. Onsager. “On the Theory of Isotope
Separation by Thermal Diffusion.” Physical Review 55 (1939): 1083–95.
Golovin, I. N., N. N. Ponomarev-Stepnoi, and L. L. Sokolovskii. “On the
275th Anniversary of the Russian Academy of Sciences.” Atomic Energy
86 (1999): 243–53.
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 300
KEMPK|KGas Centrifuges and Nuclear Weapons
301
Gordin, Michael D. Red Cloud at Dawn: Truman, Stalin, and the End of the
Atomic Monopoly. New York: Farrar, Straus and Giroux, 2009.
Gregory, W. B. “Tests of Centrifugal Pumps.” Transactions of the American
Society of Mechanical Engineers 22 (1901): 262–74.
Groth, W. E., K. Beyerle, E. Nann, and K. H. Welge. “Enrichment of Uran-
ium Isotopes by the Gas-Centrifuge Method.” In Proceedings of the
Second United Nations International Conference on the Peaceful Uses of
Atomic Energy, 439–46. New York: United Nations, 1959.
Hewlett, Richard G., and Oscar E. Anderson. A History of the United States
Atomic Energy Commission, vol. 2: The New World, 1939/1946. State Col-
lege: Pennsylvania State University Press, 1962.
Hoddeson, Lillian, Paul W. Henriksen, Roger A. Meade, and Catherine L.
Westfall. Critical Assembly: A Technical History of Los Alamos during the
Oppenheimer Years, 1943–1945. Cambridge: Cambridge University Press,
1993.
Holloway, David. Stalin and the Bomb: The Soviet Union and Atomic Energy,
1939–1956. New Haven, CT: Yale University Press, 1994.
Holweck, M. “Pompe moleculaire helicoidale.” Comptes Rendus de l'Aca-
démie des Sciences 177 (1923): 43–46.
International Atomic Energy Agency. “Implementation of the NPT Safe-
guards Agreement and Relevant Provisions of Security Council Resolu-
tions 1737 (2006), 1747 (2007) and 1803 (2008) in the Islamic Republic
of Iran.” Report by the Director General, GOV/2008/15, 26 May 2008.
IAEA, Vienna.
———. “Implementation of the NPT Safeguards Agreement and Relevant
Provisions of Security Council Resolutions 1737 (2006), 1747 (2007),
1803 (2008) and 1835 (2008) in the Islamic Republic of Iran.” Report by
the Director General, GOV/2009/35, 5 June 2009. IAEA, Vienna.
Jeffcott, H. H. “The Lateral Vibration of Loaded Shafts in the Neighbour-
hood of a Whirling Speed—The Effect of Want of Balance.” Philosoph-
ical Magazine 37 (1919): 304.
Jones, Vincent C. Manhattan, the Army and the Atomic Bomb. Washington,
DC: Center of Military History, U.S. Army, 1985.
Kelly, Cynthia C., ed. The Manhattan Project: The Birth of the Atomic Bomb
in the Words of Its Creators, Eyewitnesses, and Historians. New York:
Black Dog & Leventhal Publishers, 2009.
Kemp, R. Scott. “Gas Centrifuge Theory and Development: A Review of
U.S. Programs.” Science and Global Security 17 (2009): 2–20.
———. “Nonproliferation Strategy in the Centrifuge Age” (Ph.D. diss.,
Princeton University, 2010).
Kistemaker, J. “Hoe een nieuwe industrie onstond (deel 2): de geschiedenis
van het Nederlands Ultracentrifuge Project.” Stroom 4 (1991): 17–23.
Knight, Amy W. Beria: Stalin’s First Lieutenant, 2nd ed. Princeton, NJ:
Princeton University Press, 1993.
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 301
TECHNOLOGY AND CULTURE
APRIL
2012
VOL. 53
302
Kojevnikov, Alexei B. Stalin’s Great Science: The Times and Adventures of
Soviet Physicists. London: Imperial College Press, 2004.
Kratzer, Myron. Personal interview with author, 5 May 2011. Bonita Springs,
Florida.
Kruglov, Arkadii. The History of the Soviet Atomic Industry. London: Taylor
& Francis, 2002.
Kuhlthau, A. Robert. Personal interview with author, 27 November 2007.
Charlottesville, Virginia.
Langewiesche, William. The Atomic Bazaar: The Rise of the Nuclear Poor.
New York: Farrar, Straus and Giroux, 2007.
Lindemann, Frederick A., and Francis W. Aston.“The Possibility of Separa-
ting Isotopes.” Philosophical Magazine 37 (1919): 523–34.
“A Modest Suggestion for a Review of the Bidding.” Studies in Intelligence
14 (1970): 129 (declassified 18 April 2005).
Mulliken, Robert S. “The Separation of Isotopes by Thermal and Pressure
Diffusion.” Journal of the American Chemical Society 44 (May 1922):
1033–51.
———. “The Separation of Isotopes: Application of Systematic Fractiona-
tion to Mercury in a High-Speed Evaporation-Diffusion Apparatus.”
Journal of the American Chemical Society 45 (July 1923): 1592–1604.
Muszyn´ska, Agnieszka. Rotordynamics. Boca Raton, FL: CRC Press, 2005.
Nuclear Task Force. “Nuclear Inspection: Have and Have-Not Countries.”
University of California Radiation Laboratory, report 122-31-13B, 5
December 1955, 23.
Obeidi, Mahdi, and Kurt Pitzer. The Bomb in My Garden: The Secret of Sad-
dam’s Nuclear Mastermind. Hoboken, NJ: John Wiley & Sons, 2004.
Oleynikov, Pavel V. “German Scientists in the Soviet Atomic Project.” The
Nonproliferation Review 7 (2000): 1–30.
Pilat, Joseph, and Walter Kirchner. “The Technological Promise of Coun-
terproliferation.” Washington Quarterly 18 (1995): 153–66.
Plotkina, A. “The Development and Improvement of the Centrifuge Meth-
od to Separate Uranium Isotopes in Russia.”Atominform 6 (1996): 50–53.
Reed, B. Cameron. “Centrifugation during the Manhattan Project.” Physics
in Perspective 11 (2009): 426–41.
Rhodes, Richard. The Making of the Atomic Bomb. New York: Simon &
Schuster, 1986.
Sazykin, A. A. “Development of Gas Centrifuges for Uranium Enrichment
in the USSR.” Paper presented at the Seventh Workshop on Separation
Phenomena in Liquids and Gases, Moscow, 2000.
Shidlovsky, V. V., and G. S. Soloviov. “History and Status of Industrial Iso-
tope Separation in Russian Federation.” Paper presented at the Seventh
Workshop on Separation Phenomena in Liquids and Gases, Moscow,
2000.
Simonenko, D. L. “Kratkoe opisanie pervykh eksperimental’nykh rabot po
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 302
KEMPK|KGas Centrifuges and Nuclear Weapons
303
razdeleniyu izotopov v SSSR (1942–1948 g.g.).” In Istoriya sovetskogo
atomnogo proekta, vol. 1, edited by V. P. Vizgin, 135–85. Moscow: YanusK,
1998.
Sinev, N. M. Enriched Uranium for Atomic Weapons and Nuclear Power: On
the History of the Soviet Industrial Technology and Infrastructure to
Produce Highly-Enriched Uranium (1945–1952). Moscow: CNIIAtomIn-
form, 1991.
Steenbeck, M. Impulse und Wirkungen: Schritte auf meinem Lebensweg. Ber-
lin: Verlag der Nation, 1977.
Truman, Harry S. “Statement by the President Announcing the Use of the
A-Bomb at Hiroshima.August 6, 1945.” In Public Papers of the Presidents
of the United States: Harry S. Truman, 1945–1953, 197. Washington, DC:
U.S. Government Printing Office, 1961–66.
U.S. Atomic Energy Commission. AEC Handbook on Oak Ridge. Washing-
ton, DC: U.S. Government Printing Office, 1955.
———. “Gas Centrifuge Appraisal Report,” vol. 2: “Administration and
Planning.” Oak Ridge Operations Office, report ORO-613-2, 1963.
Vance, John M. Rotordynamics of Turbomachinery. New York: John Wiley &
Sons, 1988.
Wellerstein, Alex.“Knowledge and the Bomb: Nuclear Secrecy in the United
States, 1939–2008” (Ph.D. diss., Harvard University, 2010).
Wittner, Lawrence S. One World or None: A History of the World Nuclear
Disarmament Movement Through 1953, vol. 1: The Struggle Against the
Bomb. Palo Alto, CA: Stanford University Press, 1993.
Wood, Houston G. Ongoing personal discussions with author at Princeton
University, 2008.
Zippe, Gernot. “The Development of Short Bowl Ultracentrifuges.” Ord-
nance Research Laboratory, University of Virginia, Progress Report
ORO-202, 1 July 1959.
———. “The Development of Short Bowl Ultracentrifuges.” Division of
Engineering Physics, Research Laboratories for the Engineering Sci-
ences, University of Virginia, Progress Report ORO-216, 6 November
1959.
———. “The Development of Short Bowl Ultracentrifuges.” Division of
Engineering Physics, Research Laboratories for the Engineering Sci-
ences, University of Virginia, Final Report ORO-315, July 1960.
———. “Unclassified Spots on the History of the Modern Gas Centrifuge.”
Paper presented at the Fifth Workshop on Gases in Strong Rotation,
Charlottesville, Virginia, 5–9 June 1983.
———, with Ekkehard Kubasta. Rasende Ofenrohre in stürmischen Zeiten.
Self-published, 2008.
———, Jesse W. Beams, and A. Robert Kuhlthau. “The Development of
Short Bowl Ultracentrifuges.” Ordnance Research Laboratory, Univer-
sity of Virginia, Progress Report ORO-210, 1 December 1958.
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 303
TECHNOLOGY AND CULTURE
APRIL
2012
VOL. 53
304
Archival Collections—from the Papers of Ralph A. Lowry Archived in
the Oak Ridge National Laboratory, Oak Ridge, Tennessee
Beams, Jesse W. Letter to T. H. Johnson, 8 May 1952.
———. Letter to T. H. Johnson.“Notes on KLI-1460 by Dr. G. F. Mills on
Isotope Separation,” 7 July 1952.
Cohen, Karl. Letter to Jesse W. Beams, 1 February 1960.
Dickerson, J. H., and Ralph A. Lowry. “Notes on the Mechanical Perfor-
mance of 5-inch Diameter Gas Centrifuge Rotors,” May–July 1961.
Evans, E. C., and E. F. Babelay. “Gas Centrifuge Development, Progress
Report, November 1, 1960 through April 30, 1961.” Union Carbide
Nuclear Company, Oak Ridge, Tennessee, Report K-1481 (redacted), 1
June 1961.
———. “Gas Centrifuge Development, Progress Report, May 1, 1961
through November 30, 1961.” Union Carbide Nuclear Company, Oak
Ridge, Tennessee, Report K-1509 (redacted), 29 December 1961.
Kolstad, George A. (Division of Research, AEC) to A. Robert Kuhlthau
(Ordnance Research Laboratory, University of Virginia), 27 December
1954.
Kuhlthau, A.Robert.“Notes by A.R. Kuhlthau from Discussion with Zippe,”
1957.
———. “Check on Zippe Data,” 1960.
———. “Updated Comments by A. R. Kuhlthau about His Early Relation-
ship with Zippe,” ca.1980.
Lowry, Ralph. A. “Laboratory Notebook 7,” 7 January 1958.
———. “Laboratory Notebook 12,” 16 September 1960.
———. “Laboratory Notebook 15,” 20 February 1961.
———. Letter to E. F. Babelay, 25 September 1961.
Palmer, S. W. “Notes on the Roles of Early Members of the UCCND Gas
Centrifuge Development Group,” 30 July 1986.
Research Laboratories for the Engineering Sciences, University of Virginia.
“Proposal for the Extension of Contract AT-(40-1)-1779.” Proposal EP-
AEC-16-60S (redacted), April 1960.
———.“Gas Centrifuge Progress Summary, 1 June 1960 through 31 August
1960.” Report EP-4422-287-63S (redacted), 1961.
———. “Quarterly Progress Report for the Period 1 June 1960 to 1 Sep-
tember 1960.” Report EP-4422-134-60S (redacted), 1961.
Sickafus, E. N., R. B. Nelson, and R. A. Lowry. “The Holweck Type Molec-
ular Pump.” Research Laboratories for the Engineering Sciences, Uni-
versity of Virginia, Report EP-4422-178-61U, August 1961.
Union Carbide Nuclear Company (UCNC). “Proposal for the Develop-
ment of the Gas Centrifuge Process of Isotope Separation.” Proposal
KA-621, 11 July 1960.
———. “Gas Centrifuge Development of Process Development [sic],
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 304
KEMPK|KGas Centrifuges and Nuclear Weapons
305
Quarterly Report for Third Fiscal Quarter, Jan 1, 1961 – March 31,
1961.” Report K-1502 (redacted), 24 July 1961.
———. “Gas Centrifuge Development of Process Development [sic],
Quarterly Report for First Fiscal Quarter, July 1, 1961 – September 30,
1961.” Report K-1504 (redacted), 30 January 1962.
———. “Gas Centrifuge Development of Process Development [sic],
Quarterly Report for Second Fiscal Quarter, October 1, 1961 – Decem-
ber 31, 1961.” Report K-1505 (redacted), 27 April 1962.
———. “Gas Centrifuge Development of Process Development [sic],
Quarterly Report for Third Fiscal Quarter, Jan 1, 1962 – March 31,
1962.” Report K-1506 (redacted), 16 August 1962.
Waters, Dean A. “A Vibration Study of Subcritical Gas Centrifuges.” UCNC
Report K-1485, 3 October 1961.
04_TEC53.2kemp 272–305:03_49.3dobraszczyk 568– 4/30/12 10:55 AM Page 305
