The Last Man Who Knew Everything

by David N. Schwartz

  Given the central importance of Fermi’s 1925 work on statistical mechanics to Chandrasekhar’s entire career, the only surprising aspect of the relationship between the two is that it seems to have started so late. Chandrasekhar arrived at the University of Chicago in 1937 but waited until many years later to send Fermi a rather formal letter inviting him to visit the observatory some one hundred miles northwest of Chicago on the shore of Lake Geneva, Wisconsin, in November 1947. Perhaps the delay reflected Fermi’s hectic schedule during the war or perhaps it reflected Chandrasekhar’s shyness. He might simply have been waiting respectfully until Fermi had settled into the new institute. Fermi responded positively and met with Chandrasekhar. The two of them hit it off well, the letters between them became increasingly less formal, and in the summer of 1949 Chandrasekhar invited both Enrico and Laura to spend time with him and his wife at their home near the observatory.

  In the fall of 1953, the discussions between Fermi and Chandrasekhar became more regular. Chandrasekhar would spend two days a week on the Chicago campus, and the two would meet to discuss cosmic rays, galactic magnetic fields, and the like over lunch at the faculty club. In discussions with Chandrasekhar, Fermi extended the work he did in 1948, taking account of the spiral arms of the Milky Way galaxy and the magnetic fields created by them. The two men published several papers on the subject together during this period.

  Fermi had no formal training as an astrophysicist, but Chandrasekhar likened him to a musician who, confronted with a new piece of sheet music, could perform it the first time through with the brilliance of an artist. For Chandrasekhar, the experience of working closely with Fermi was one of the highlights of his life. For his part, Fermi was willing to confide in his new friend, especially when it came to explaining how he thought about problems.

  In retrospect, Fermi and Chandrasekhar were only partly correct about how cosmic rays develop such high energies. Astrophysicists still believe that magnetic fields are responsible for some of the high-energy particles that collide with Earth, but these are not necessarily free-floating interstellar fields. Rather they are associated with supernovae and with some highly unusual objects that were unknown in Fermi’s day, objects such as quasars, pulsars, and other highly dense, rotating objects with strong magnetic fields. Also, the universe is far larger than they knew then, with far more potential sources of radiation. Yet with all that was unknown at the time, Fermi’s method of studying the subject and coming to tentative insights demonstrates the power of his approach.

  THOSE LIKE BRUNO ROSSI, WHO HAD BEEN STUDYING COSMIC RAYS, knew that particles like protons, pions, and muons were continuously colliding with the earth’s atmosphere, producing showers of other subatomic particles. The collision of pions and their subsequent decay into other particles was a subject of some interest to Fermi in the first part of 1947. Fortunately, he had a new graduate student who was not doing particularly well in his theoretical physics thesis topic and who was willing, indeed eager, to follow Fermi’s suggestion and conduct an experiment to study muon decay in the atmosphere. In the process, the student unwittingly demonstrated for the first time the existence of one of the fundamental forces of nature. His name was Jack Steinberger.

  Steinberger, a German Jew who arrived in Chicago with his parents before the war, had not only been having trouble with his thesis but had trouble with the physics program from the beginning. He failed the basic exams required prior to beginning work on his doctoral thesis and was, thanks to the generosity of department chair Walter Zacharaisen, the only student ever to be given a second chance. Fermi clearly liked the young, strikingly handsome Steinberger. He asked Steinberger to be his teaching assistant for a course in elementary physics that Fermi taught in the fall of 1946. Fermi also agreed to be Steinberger’s thesis adviser after his student finally passed his basic exams.

  Fermi could see that Steinberger was getting discouraged with his theory dissertation and suggested gently that perhaps an experimental project would be more to his liking. Fermi’s old friend Bruno Rossi and others had been studying the way cosmic muons, created by cosmic pion decay, themselves decayed into electrons and had discovered far fewer electrons than predicted. The rate was lower than expected by a factor of two. Sensing that it might be an interesting story, Fermi suggested that Steinberger study the spectrum of electron energies resulting from these decays.

  It is a rare PhD dissertation that makes an important scientific contribution to the field, but largely because of Fermi’s sixth sense for important questions and also because of the unusual results of Steinberger’s experiment, this dissertation research was an exception to the rule.

  Fermi lent Steinberger an assistant, and together they made some eighty Geiger counter detectors configured appropriately to capture electron tracks from cosmic muon decays. The first phase of the experiment was completed at sea level and later repeated in 1948 at the top of Mount Evans in Colorado, at an elevation of 14,271 feet, to enhance the statistical significance of the results. Steinberger discovered that the energy spectrum of electrons emitted from cosmic-ray muon decay was continuous and even at the highest energies was not sufficient to account for nearly half the energy that the muon’s decay should produce. Steinberger analyzed the data carefully and came to the conclusion that two neutral particles, “probably neutrinos,” accompany each electron emitted from the muon decay. These two particles carried off the missing energy.

  Neither Steinberger nor Fermi realized the most important implication of this result. Steinberger says that he himself was not “clever enough.” It is difficult to see why Fermi missed it. Perhaps, as Steinberger suggests, “new ideas are not always easy to accept, sometimes even by the brightest and most open of people such as Fermi.” In any case, several others did, including three University of Chicago graduate students, Tsung-Dao Lee, Chen-Ning Yang, and Marshall Rosenbluth. They proposed that what Steinberger discovered was a more general extension of the “Fermi interaction” underlying beta decay. They suggested that there was a “universal” force—alongside gravity, electromagnetism, and the “strong” force holding the nucleus of an atom together—that produced changes in particles and that the neutrino always appeared in these processes as a way to account for the energy that was not imparted to the other particles created in these processes. This universal force came to be known as the “weak” force, and the creation and/or destruction of neutrinos came to be called the “weak” interaction, because the force responsible for this can only be felt at the closest of distances.

  Steinberger recalls that Fermi was extraordinarily generous with his time, organizing the logistics and funding for Steinberger, but never interfered with the actual conduct of the experiment itself, allowing the young physicist to make his own mistakes along the way. Fermi explained to Steinberger, as he would to Maria Mayer at about the same time, that if Fermi added his own name to the paper reporting the results of the experiment, everyone would think that Fermi had done all the work, which would be bad for Steinberger.

  CONFERENCES CONTINUED TO REVEAL AND DISCUSS IMPORTANT NEW developments in postwar physics, but the action shifted to the United States, where a series of major conferences organized under the auspices of the National Academy of Sciences effectively replicated the prewar Solvay conferences. The first was scheduled for early June 1947 at Shelter Island, situated between the north and south forks of Long Island, New York. Fermi received an invitation and was eager to attend. The agenda included discussion of some of the most interesting new developments in physics: Willis Lamb on a strange anomaly he had discovered while measuring the energy of the two possible quantum states of the hydrogen atom, an anomaly which would soon be known as the Lamb shift; Rabi on his precise experimental measurement of the magnetic moment of the electron, a value that Dirac’s theory of quantum electrodynamics could not compute; Robert Marshak on the Yukawa meson, whose experimental observation a few months later would vindicate Marshak’s speculations; and Feynman on an informal, p
reliminary presentation of his work in quantum electrodynamics and his use of a strange new analytical tool based on graphic diagrams of interactions. The stellar attendance list included among others Bethe, von Neumann, Oppenheimer, Rossi, Teller, Uhlenbeck, Weisskopf, and Wheeler. Fermi would have been in his element.

  Fermi never made it to the conference. Passing through Baltimore, he noticed that his vision was blurred. Understandably alarmed, he decided to have a doctor examine him. The cause was a torn retina, which would take the better part of a year to heal completely. During this time, his friends would often notice him counting the ridges of his fingerprints or carefully holding a pencil in front of his eyes, trying to focus on the edge of the pencil to make sure his eye was returning to normal.

  A second conference, at a resort in the Pocono Mountains of Pennsylvania, took place in late March 1948. By this time Fermi’s eyesight had recovered sufficiently for him to attend. In addition to the participants who were at Shelter Island, several other old Fermi friends attended, including Niels and Aage Bohr, Dirac, Wigner, and Wentzel. The main focus of the conference was quantum electrodynamics (QED), and two presentations on the subject made history.

  For decades the niggling problem inherent in Dirac’s version of QED defied repeated attempts, including those by Fermi, to solve it. The completion of the magnetic moment of the election converged not on a finite value, as one would expect, but instead diverged to infinity. Between 1946 and 1948, three young physicists independently gave the problem another shot, and this time all of them succeeded. It was a great achievement, one of the greatest of twentieth-century physics, and put QED in an almost uniquely effective category of physical theory. Two of these physicists were invited to the Pocono conference—Feynman and a thirty-year-old Harvard professor named Julian Schwinger. (The third, a Japanese theorist named Sin-Itiro Tomonaga, independently developed a solution similar to Schwinger’s and would certainly have been invited to present at the conference if the organizers had known about his work.)

  Schwinger and Feynman, though the same age, could not have been more different, as individuals or as physicists. Feynman was a gregarious and irrepressible showman; Schwinger was a quiet introvert, with no interest at all in entertaining his audience. As a physicist, Feynman’s presentation reflected an informal, intuitive approach that depended upon his diagrams, which eventually found universal acceptance. Schwinger was more of a formalist, relying on equation after equation, carefully sequenced, to get to his solution. His presentation was so long and tedious that in the end only two physicists lasted through the entire lecture: Fermi and Bethe. Fermi took it as a badge of honor that he stayed until the last.

  Although he had great respect for Feynman throughout his life, Fermi clearly preferred the Schwinger approach, based squarely on the traditional quantum field theory Fermi knew so well. Feynman’s approach was anything but traditional, and the graphic tools Feynman used did not appeal to Fermi, at least not immediately. In 1951, Fermi and Franck jointly nominated Schwinger for a Nobel Prize. They cited Feynman’s and Tomonaga’s QED work, but they believed that “the greatest contribution was made by Schwinger.” Feynman never knew this. He continued to correspond with Fermi on a wide variety of physics problems and visited Fermi from time to time at Chicago. Feynman had enormous respect for Fermi, and part of his eagerness to bounce ideas off Fermi may have been an unconscious effort to gain the validation as a physicist that he sought but did not receive at the Pocono conference.

  Why is it that Feynman, Schwinger, and Tomonaga succeeded in 1947 where Fermi (and others) failed in the early 1930s? Were they just better theorists? Given what we know about Fermi’s formal abilities in mathematical physics and his ability to crack profound problems in quantum field theory, such a conclusion seems, on the surface at least, too facile. Perhaps a better explanation is rooted in Peierls’s observation that Fermi was attracted to problems in which the mathematics was relatively simple. Peierls’s implication is that Fermi was an impatient theorist, that when a problem did not quickly succumb to his intellect he quickly lost interest in it. Another possibility: Fermi was an intensely practical physicist, not inclined to put more effort into a problem than he felt it was worth. It is entirely possible that he knew that the first approximation of the magnetic moment of the electron provided by Dirac’s theory was good enough for most practical purposes and that finding a way to calculate the value out to five or six decimal places was simply not worth his time and energy. As Fermi told his daughter Nella in another context, “never make something more accurate than absolutely necessary.” After the war it was not a problem to which he devoted any time. This is not to deny the historic achievement of these three, who shared the Nobel Prize for their work in 1965. Yet for Fermi, it may not have been that interesting a problem or perhaps he knew that by 1947 he didn’t have the youthful brilliance and intellectual stamina required to crack it and was happy to give others a chance. For whatever reason, he was clearly more concerned with other scientific matters.

  ACADEMIC SUMMERS ONCE AGAIN BECAME PERIODS OF RELATIVE freedom, and Fermi took advantage of them as he had before the war. Most summers he spent six or eight weeks at Los Alamos, doing research on various projects, mostly classified, and spending time with friends, old and new. Soon after the war, the McMahon Act of 1946 placed the control of all atomic energy research under US civilian government monopoly, focused on a handful of national laboratories, including Los Alamos. Congress initially considered a more restrictive law that would have kept research under military control. Fermi, eager to present a united front with Oppenheimer, supported this along with Compton and Lawrence. However, other Manhattan Project scientists were outraged by continued military control of nuclear research and particularly unhappy with the four scientists who gave their support. Herb Anderson delivered a stinging critique of the group that included his old friend and mentor:

  I must confess my confidence in our own leaders Oppenheimer, Lawrence, Compton, and Fermi, all members of the Scientific Panel,… who enjoined us to have faith in them and not influence this legislation, is shaken. I believe that these worthy men were duped—that they never had a chance to see this bill. Let us beware of any breach of our rights as men and citizens. The war is won, let us be free again!

  Widespread opposition to the bill ultimately led its sponsors to withdraw it and replace it with a 1946 bill, sponsored by Connecticut senator Brian McMahon, putting the nuclear program under civilian control. It created the Atomic Energy Commission (AEC), whose first head was former director of the Tennessee Valley Authority David Lilienthal. It established a General Advisory Committee (GAC) of science and technology experts to advise the AEC. Oppenheimer was named the GAC’s chairman and Fermi was selected for a four-year term. Under the AEC, new labs at Los Alamos, Argonne, Oak Ridge, Hanford, and Berkeley continued working to refine the country’s atomic arsenal and pursued basic research on a variety of scientific fronts. The new director of Los Alamos, Norris Bradbury, was eager to bring Fermi back to the mesa, and Fermi was delighted to return, family in tow.

  FIGURE 21.3. Summers at Los Alamos after the war were spent with old friends. Here is Fermi, third from left, on an afternoon jaunt with Hans Bethe, L. D. P. King’s son Nick (behind the wheel), and Edward Teller’s son Paul. Courtesy of Los Alamos National Laboratory.

  Upgrading from wartime austerity, the Fermi family settled into one of the nicer homes on Bathtub Row and Enrico cycled to work, while Laura and the children socialized and pursued summer activities. He worked on a variety of classified projects, but two unclassified projects were of particular interest.

  First, he became enamored with computers. Before electronic computers were readily available, he designed analog computers that could help him with his calculations. One such computer, built in Chicago with the help of graduate student Richard Garwin, was capable of solving Schrödinger equations. Another one, built at Los Alamos with the help of his old water boiler colleague L. D. P. King, used
Monte Carlo methods to trace the path of a neutron through matter, simulating the paper and pen calculations Fermi did on neutron diffusion before the war. It was affectionately dubbed “Fermiac,” after the early electronic computer ENIAC.

  By 1947–1948, von Neumann developed one of the first programmable electronic computers and it became the focus of attention at Los Alamos. The father of computational physics, a young Manhattan Project veteran named Nicholas Metropolis, used von Neumann’s machine to explore how physics equations could be programmed into a computer to simulate physical processes and predict the outcome of experiments. With von Neumann and Stanislaw Ulam, Metropolis invented the modern computerized “Monte Carlo” method of simulating stochastic physical processes. Fermi jumped at the chance to put his own equations into the computer, working with Metropolis on a wide variety of studies, including what would happen if a high-energy pion hit the nucleus of a relatively simple hydrogen atom. Several years later, these simulations allowed Fermi to compare his theoretical predictions with actual experimental results when the Chicago cyclotron went live. Later, in 1953, Fermi worked with Ulam and another computational physicist, John Pasta, to study the problem that absorbed him in his 1918 application to the Scuola Normale Superiore. They programmed the equations of a vibrating string into the Los Alamos computer and simulated its behavior. The resulting paper became an early and crucial contribution to chaos theory by showing that the string would return to a specified state at regular intervals, which was inconsistent with its expected ergodicity. In this respect it was a direct lineal ancestor of Fermi’s very early concerns about ergodic systems and their behavior.