"Golden Books" - The Beginnings of Fermilab
The Beginnings of Fermilab
Lillian H. Hoddeson
Fermi National Accelerator Laboratory
Operated by Universities Research Association, Inc.
Under Contract with the United States Department of Energy
Lillian Hoddeson has been Fermilab's historian since 1978. Born in the Bronx in 1940, she attended the infamous physics prep school, the Bronx High School of Science, eventually receiving her doctorate in physics from Columbia University. However, she decided that she could make a more distinctive contribution by studying the history of physics, concentrating on twentieth century developments in particle physics and accelerators, solid-state physics, large laboratories, and atomic weapons. She has written books and articles on all of these subjects.
While leading the history and archives program for Fermilab's Directors Office, concurrently with teaching and research in the departments of physics and history at the University of Illinois at Urbana-Champaign, Lillian and collaborators organized three international symposia on the history of particle physics and edited their proceedings. She has served on panels and committees for the American Physical Society, American Institute for Physics, and National Science Foundation, and has consulted on the history of physics for the American Institute of Physics, Bell Laboratories, and Los Alamos National Laboratory. She has a daughter Carol and a son Michael.
This selection is abstracted from Lillian's 1983 article, "Establishing KEK in Japan and Fermilab in the U. S.: Internationalism, Nationalism, and High Energy Accelerators," comparing Fermilab's early history with that of our Japanese counterpart, published in Social Studies of Science 13, 1-48.
The Beginnings of Fermilab
Viewpoint of an Historian
In the wake of technical and institutional developments spawned by the Second World War, the art and science of accelerator building was blossoming into a strongly competitive, independent branch of physics. Innovations appeared by the dozens. To list but a few of the more important ones: phase stability; alternating gradient (AG) or "strong" focusing; separated-function magnets; beam stacking; colliding beams; the rapid-cycling high-intensity synchrotron; and the fixed-field alternating gradient (FFAG) accelerator.
In the early 1950s, the principle American accelerator efforts were still localized on the East and West coasts; the Brookhaven Cosmotron and Berkeley Bevatron, both proton synchrotrons, were the largest accelerators in the world, as well as the first to be funded by the U.S. Atomic Energy Commission (AEC). A pioneering 32-MeV proton linear accelerator-the prototype for a generation of subsequent linear accelerators-was built at Berkeley, using wartime surplus radar. By the middle of the decade, however, these East and West coast efforts were being challenged by accelerator developments elsewhere in the country, for example by those of the Midwestern Universities Research Association (MURA), formed soon after the Cosmotron had been completed in 1952, which hoped to design the next large United States accelerator facility. Reasonably, they felt, this next accelerator should be located in the Midwest. By this time, the various American accelerator schools were differentiating one from another in their approach to design.
The Fermilab machine's conceptual root grew out of the discussions at a MURA summer study meeting in 1959, a meeting aimed at generating support for MURA and reconsidering the FFAG design in relation to all the existing schemes of achieving high energy or high intensity. At that time it was generally felt that the only practical way to produce energies in the several-hundred-GeV range was by colliding accelerated beams. Furthermore, it was believed that fixed-target machines of very high energy would be exorbitantly expensive (if even feasible technically)-and perhaps not useful for physics anyway, because above approximately 5 GeV all the existing schemes for identifying particles were suspect. (The fear was that all particles would look alike, being confined to a narrow forward-moving cone.) At that meeting, Matthew Sands, an iconoclastic participant from the California Institute of Technology (Caltech), became challenged by the problem of designing a reasonable cost fixed-target machine aimed at approximately 300 GeV. He reinvented the concept, suggested several years earlier by Robert R. Wilson and others (including F. Heynman and Lee Teng), of forming a cascade of accelerators, injecting an accelerated beam from one machine to another. By first accelerating the particles up to a high energy in a "booster" synchrotron and then, with a reasonably high injection field, feeding the beam into a main synchrotron, Sands hoped to avoid the use of very large (and, therefore, very costly) magnets in the highest energy machine. Beams of particles occupy increasingly smaller transverse space as they experience acceleration. By cascading accelerators, the final and largest ring would then require smaller transverse dimensions.
Although it was thought then that one could not control such a large system or use magnets as small as Sands specified, working out details with a subgroup of the MURA study (which included Ernest Courant and M. Hildred Blewett from Brookhaven, and Alvin Tollestrup from Caltech), Sands showed mathematically that the magnet aperture in the main ring could be but a few square centimeters in size. Optimization of parameters gave the result that to achieve 300 GeV most efficiently, one should inject into the main ring from a 10GeV range "rapid cycling," booster synchrotron. The high repetition rate of such a booster would enable a high intensity to be achieved.
Most of the participants at the MURA summer study did not take Sands' proposal seriously, but he and Tollestrup continued to work on the idea at Caltech, also involving Robert Walker in the project. Since Caltech judged building such a machine to be a project too large for them to support alone, a sponsoring group was formed, the Western Accelerator Group (WAG), which included physicists from Caltech, the University of California at both Los Angeles and San Diego, and the University of Southern California. Berkeley declined the offer to join WAG, for in the late fifties, researchers there (for instance, David Judd and Lloyd Smith) had been working on their own concept for a very-high-energy machine based on a proposal of Nicholas Christofilos. In April 1961, WAG submitted its proposal to the AEC. One of those who appreciated WAG's design was Wilson at Cornell, whose earlier suggestion of cascade injection had stimulated the Sands design. Several years later Wilson would build Fermilab on this model. Presciently, he wrote to Sands on April 25, 1961: "I have been watching your efforts with the 300-GeV machine with open-mouthed admiration. It seems to me that you are working on the right problem and at the right time, and I am sure something will come of it all."
Meanwhile, interest in a several-hundred-GeV machine had been mounting in other parts of the U.S. In August 1960, Wilson organized an unofficial conference in Rochester, New York, at which approximately 30 physicists who were attending the concurrent Rochester Conference on High Energy Physics took part in "intensive discussion of. . . the desirability of super energy from the point of the theory of particles. . . [and] the experimental practicability of constructing and using ultra-high-energy machines." As Wilson summarized in his own style:
"It was generally agreed that for, say, $100 million-or at most $200 million-it would be feasible to push the design of a conventional alternating-gradient proton synchrotron to 100 GeV or even higher and that this might also cover the first round of experiments. With the same reasoning, but pushing the kind of tolerances that must be held, we could even think of attaining 1000 GeV and at a cost of less than $1 billion-really a bargain, of course."
Further support for such projections came at a meeting in September 1960 at the American Institute of Physics in New York, and at the 1961 International Particle Accelerator Conference at Brookhaven. In February 1962, Brookhaven submitted a proposal to the AEC for a 300- to 1000-GeV design study. In February and December of 1962, Berkeley submitted proposals for study of machines in the 100- to 300- GeV range. WAG's 300-GeV proposal was now in serious competition with proposals from the established accelerator laboratories.
The AEC, having also to evaluate proposals for other large machines, including one by Cornell to upgrade its machine and one by MURA to build a 10-GeV FFAG, found itself in need of advice. For the first time in its history, the funding was insufficient to support all pending accelerator proposals. From this point on, American high-energy physicists would be spending more and more time on panels to discuss funding. Extensive participation by the U.S. Congress Izad recently begun with hearings in 1959-60 over the issue of supporting Stanford' s linear accelerator, the first machine with a budget in the $100-million range.
Most influential of the American accelerator panels in this period was that headed by Norman Ramsey during 1962 and 1963. After extended discussion, this panel ranked the proposals and, in a report in April 1963, suggested 13 steps to be taken in order. The first was that Berkeley, rather than Brookhaven or Caltech, construct a proton accelerator of approximately 200 GeV. This machine eventually became Fermilab. The next three steps were: that Brookhaven construct storage rings "after a suitable study"; that design studies be conducted at Brookhaven for a 600- to 1000-GeV national accelerator; and that MURA, in fiscal year 1965, construct "a super-current accelerator without permitting this to delay the steps toward high energy . . ." WAGS's proposal was not mentioned. Brookhaven and Berkeley were favored for the very-high-energy machines because they were the most experienced accelerator laboratories. By ranking the MURA machine fourth, the panel effectively phased out this machine, a move that would later enter into the selection of the site of the 200 GeV accelerator.
Two other features of the Ramsey Panel's report are notable. First, the recommendation emphasized that studies for new high-energy facilities "should be permitted to proceed to greater detail with explicit authorization so that ideas can be explored conclusively without implying any commitment to proceed." Thus, accelerator development advanced into the era of the "design study," in which groups of physicists would be authorized to prepare detailed designs over a period of several years, without any commitment to build. Secondly, the panel stressed that future high-energy laboratories be nationwide rather than regional facilities, having a strong users' representation as well as in-house research staff.
An informal but influential paper, prepared in June 1963 by Leon Lederman of Columbia University, then participating in the committee headed by Myron L. Good and appointed to review the Ramsey Panel's report, defined the concept of the "Truly National Laboratory," or TNL a laboratory whose ultimate governing body, to which even the director would be responsible, would be a nationally represented committee. The users' group at the TNL would be "at home and loved." Not only, Lederman argued, should users have the right of access to the machine, ancillary equipment, and any specialized services that are offered, but also (1) laboratory and office space on site; (2) a "substantial" support budget to supplement their own grants; (3) strong representation on the scheduling committee; and (4) an active users' advisory committee. He suggested that the site be selected with a view towards "ease in airport-to-site transportation, housing and school facilities, and general pleasantness." This TNL concept would be put into operation four years later in the design of Fermilab.
After the Good Committee endorsed the Ramsey Panel's recommendations, the AEC appropriated money for Berkeley, under the direction of Edward Lofgren, to conduct a detailed study to design a 200 GeV accelerator. And two years later, in June 1965, the design study appeared. It described, in two thick blue books, a four-accelerator cascade in the spirit of the Sands proposal, but differing substantially from Sands' concept in technical features, most notably its size. In the Berkeley design, the magnet apertures were comparatively huge, contributing to the total budget of over $340 million.
It was a poor time for Berkeley to present such an expensive proposal, for Congress was just then beginning to feel that high-energy physics was over-supported, and that too large a proportion of funds was going to California. Furthermore, non-Berkeley physicists were complaining that in the past they had not been granted adequate access to Berkeley's machines. In this context, in the fall and winter of 1965, Wilson dramatically entered the story of the 200-GeV accelerator. He had had an opportunity during the previous summer to study the Berkeley design, as presented by Edwin McMillan at a meeting in Frascati, Italy. Feeling strongly that the Berkeley design was too conservative, and thus much too expensive, Wilson wrote a series of critical papers. By December, he had drawn up an alternative proposal for a 200-GeV machine at a cost of only $50 million, estimating only $100 million to achieve 600 to 1000 GeV. He based his estimates on economizing features used in the Cornell electron synchrotron-for example, small magnets and austere experimental facilities. Another alternative, suggested by Samuel Devons of Columbia University, was to add a further level of acceleration to the Brookhaven machine, using the AGS as injector. While the Berkeley physicists tended to dismiss the economizing suggestions of both Wilson and Devons, the AEC did not, and announced a cost ceiling of $240 million, so that Berkeley had to prepare a "reduced scope" design.
The debates in 1965 further focused on the location of the new laboratory. While Berkeley had assumed throughout that the site would be in California, physicists and politicians in other states actively began to question this assumption. In April 1965, after receiving Colorado's independent site proposal, the AEC advertised for other proposals. One hundred and twenty-six were received, suggesting over 200 sites, with one or more from each of 46 states. By September 1965, the AEC had reduced the number of proposals to 85, and in March 1966, with the help of a National Academy of Sciences site evaluation committee headed by Emanual Piore of IBM, only six remained. It is widely believed that political agreements entered into the final selection of the site at Weston, Illinois, about 30 miles west of the center of Chicago, a choice made in December 1966. With this choice, President Lyndon Johnson could repay a debt to the Midwest incurred by the closing of MURA in 1965, at the same time obligating the Illinois senator Everett Dirksen to support Civil Rights legislation. This plausible conjecture has been contradicted by, among others, Glenn Seaborg, the Chairman of the AEC.
Somewhat earlier, Frederick Seitz, President of the National Academy of Sciences, had taken the initiative of organizing a national university-supported organization, modeled after Brookhaven's AUI, and named the Universities Research Association (URA), to build and operate the new accelerator laboratory. In June 1965, the URA consortium, composed originally of 34, and later of over 50 and then more than 60, universities broadly distributed throughout the United States and later Canada, was incorporated. Ramsey was selected as its President. The URA's first job was to choose a director for the new laboratory. It was initially intended to have the position divided into a physics and an accelerator director. The first offer, that of accelerator director, went to Lofgren, who had been head of the Berkeley design project. But Lofgren, apparently supporting Berkeley's hope that the laboratory be in California, turned the offer down on the grounds that the Illinois site was unsuitable and the $240-million budget impossibly low. Then on February 6, 1967, the URA formally offered Wilson the combined position of accelerator and laboratory director. Wilson was appointed on March 7, 1967.
Operating from his home base in Cornell, Wilson spent the remainder of the academic year 1966-67 on staffing, designing, planning conferences, and arranging for an engineering firm to take on the construction. Staffing was somewhat hindered by the fact that the Illinois site-6800 acres of totally flat cornfield-seemed an unappealing place to live. Staffing was aided, however, by the fact that MURA and the Cambridge Electron Accelerator were then both at the point of closing down. To emphasize his intention to make the facility "truly national," as discussed by Lederman, Wilson named it "The National Accelerator Laboratory," NAL. Seven years later, in May 1974, NAL would be formally renamed Fermi National Accelerator Laboratory, or Fermilab for short, in honor of Enrico Fermi.
Wilson made effective use of the centrality of the Illinois site, bringing in people from both coasts in an almost continuous series of meetings in order to build consensus and confidence within the physicists' community as well as with the AEC and the contractors. Since Illinois was having local difficulties buying the land that was to be turned over to the Federal government, the new laboratory was not able to move to Weston until fall 1968. The workshops Wilson held in the summer of 1967 to design the laboratory, attended by participants from various parts of the U.S., supported by their home institutions, were held in temporary offices in Oak Brook, a suburb of Chicago. The conferees at Oak Brook chose basic parameters, such as the radius of the main synchrotron ring, and decided where on the site to place particular components. The workshops also gave Wilson and the conferees a chance to look each other over as potential staff and boss. At the end of the summer, approximately half those attending the workshop joined Wilson's staff.
At the Oak Brook meetings, some of the physicists argued that an added ring of superconducting magnets, installed in the Main Ring tunnel, could be used as a beam stretcher, to lengthen the time over which beam feeds out to experiments, or to store the accelerated particles that could then collide against other particles emerging from the primary ring. But most of those who met at Oak Brook considered these concepts to be beyond immediate technological feasibility; designing and building a 200-GeV non-superconducting machine was the immediate job at hand. Following these meetings, Wilson issued an informal edict prohibiting active work on a superconducting accelerator until the main accelerator was functioning. Nevertheless, he insisted that a space "be left free just below the magnets of the NAL proton synchrotron so that a second ring of superconducting magnets could be installed . . . then the energy of the protons that have been accelerated in the ring will be doubled to 1000 BeV." Richard Lundy, who worked on building the Main Ring magnets, recalls that "Bob [Wilson] did enforce the idea that space be left clear ... [although] it was never exactly obvious what would go in there."
The design report for NAL, completed during the fall of 1967 and issued in January 1968, described a cascade machine quite similar to that proposed by Sands in 1959, but with some features of the Berkeley design. Many innovations reduced costs: small "H-design" magnets with minimal enclosures and a relatively small Main Ring tunnel, separated-function magnets for bending and focusing in the Main Ring, modular equipment in the Main Ring, a single emergent beam split after extraction, newly developed solid-state rectifiers (instead of traditional flywheel generators) tying the magnets directly to AC power lines, an electrostatic septum, a Main Ring tunnel built directly on glacial clay, and simple stands rather than expensive girder supports for the magnets. The design also included a built-in option to raise the energy of the conventional machine to 400 GeV. By mid-April, Congress had passed, and Johnson had signed, the bill authorizing the project at $250 million.
The Linac group was the first to begin work at Weston in 1967-68. The rest of the staff moved there in October 1968. Then a frantic three-year period of actual construction began in December 1968 with the Linac ground-breaking. The emphasis in this period was on economy and speed; Wilson and his Deputy Director, Edwin Goldwasser, kept setting tight schedules and trying to motivate the staff to beat these schedules in order to save labor costs. Experimental facilities (including a Meson area, a Neutrino area, and a Proton area) were planned by a national group at summer studies held in 1968 and 1969 in Aspen, Colorado, and in 1970 at NAL. When the first 200-GeV beam passed through the Main Ring in March 1972, the NAL accelerator was the highest energy accelerator in the world.
One year earlier, in March 1971, Wilson had begun his campaign to reserve this distinction for Fermilab during the 1980s. He explained to the joint Committee on Atomic Energy how, using superconducting magnets, the energy of the NAL machine could be doubled-the accelerating scheme that would later be called the Energy Doubler, or TEVATRON: "The idea is to take the protons out of the present magnet ring and then inject them into the new ring of superconducting magnets piggy back upon the other." He fantasized on the implications: "One could install one of these rings after another, taking the beam from one to the next, doubling the energy each time." A short proposal in February 1972, by William Fowler and Paul Reardon, set the project in motion. Initially a poorly supported, loosely organized, back-burner effort of a handful of researchers surrounding Wilson, this project would grow over the next ten years into a well-funded, well-orchestrated, large-scale effort, Fermilab's first priority.
The lack of adequate funding for the Doubler was the immediate reason for Wilson's resignation in 1978. The machine would be completed during the tenure of Fermilab's second director, Leon Lederman, after many financial and technical hurdles had been overcome. On July 3, 1983, the milestone was met of achieving the first acceleration of beam in the Doubler to 512 GeV. As J. Richie Orr, then Head of the Fermilab Accelerator Division, recalls, the event was "a pleasant surprise. The machine was a lot better than we thought it would be. . ." Experiments using the Doubler began in October and a record of 900 GeV was reached on February 16, 1984. The Fermilab Accelerator-now the Doubler was again the highest energy accelerator in the world.
Sue Grommes ● Adrienne Kolb ● Rocky Kolb ● Bruce Chrisman
John Peoples ● Robert R. Wilson ● Leon M. Lederman
S. Chandrasekhar ● Leonard Euler's Illustrator ● Angela Gonzales
Pablo Picasso ● East Indian Mythology ● 5th Century B.C. Athenian Vase
5th Century B.C. Greek Mythology ● Leonardo da Vinci
Painters and Potters ● Hans Bethe ● Fermilab Photography ● Al Johnson
John Keats ● Emily Dickinson ● Jean Lemke ● Marcel Proust
Chocolate ● Lillian Hoddeson ● William Shakespeare
19th Century Illustration Depicting a Mechanical Universe
This publication was selected from the treasures of the Fermilab Archives.