Notice: We are in the process of migrating Oral History Interview metadata to this new version of our website.
During this migration, the following fields associated with interviews may be incomplete: Institutions, Additional Persons, and Subjects. Our Browse Subjects feature is also affected by this migration.
Please contact [email protected] with any feedback.
This transcript may not be quoted, reproduced or redistributed in whole or in part by any means except with the written permission of the American Institute of Physics.
This transcript is based on a tape-recorded interview deposited at the Center for History of Physics of the American Institute of Physics. The AIP's interviews have generally been transcribed from tape, edited by the interviewer for clarity, and then further edited by the interviewee. If this interview is important to you, you should consult earlier versions of the transcript or listen to the original tape. For many interviews, the AIP retains substantial files with further information about the interviewee and the interview itself. Please contact us for information about accessing these materials.
Please bear in mind that: 1) This material is a transcript of the spoken word rather than a literary product; 2) An interview must be read with the awareness that different people's memories about an event will often differ, and that memories can change with time for many reasons including subsequent experiences, interactions with others, and one's feelings about an event. Disclaimer: This transcript was scanned from a typescript, introducing occasional spelling errors. The original typescript is available.
In footnotes or endnotes please cite AIP interviews like this:
Interview of Richard Garwin by W. Patrick McCray on 2001 June 7,
Niels Bohr Library & Archives, American Institute of Physics,
College Park, MD USA,
For multiple citations, "AIP" is the preferred abbreviation for the location.
Focused interview deals with Garwin's childhood and schooling at Case and University of Chicago; his time at Chicago and interaction with Fermi. Majority of interview is on Garwin's work in designing the H-bomb and consulting work at Los Alamos in 1950-51. Recollections interspersed with technical details of design and apparatus. Discussion of recent 'discovery' of interview with Edward Teller in 1979 by George Keyworth that gives credit for H-bomb design to Garwin; Teller and Ulam's relation. Last part of the interview deals with Garwin's work in conjunction with Leon Lederman to confirm the non-conservation of parity in late 1956-early 1957.
Finn Aaserud had interviewed you quite extensively before in the late 1980s. These interviews are now open. So I would like to focus on topics he didn't cover and more recent issues. Also, spend some time talking about the recent discussion about the history of the hydrogen bomb that's come up after the publication of the New York Times article. [April 24, 2001 "Why Build the H-bomb? Debate Revives".]
I had a long interview yesterday with Dan Charles who's writing an article for New Scientist. [Published in the August 3, 2001 issue.]
He tape-recorded it, so you could probably get a copy of that. He's a sensible person.
I thought since Finn didn't cover it too extensively it would be useful for historians in the future to have a better understanding of some of your childhood experiences and early undergraduate experiences as a way of setting the stage for the later work that you did. I was wondering if we could talk about that to some degree? For example, I'm curious, because it didn't come out in Finn's interview quite so much of how you became interested in science in the first place and how you chose that particular path. I thought we could begin there.
I mentioned that my father had graduated as an electrical engineer from what's now Case Western Reserve University, which was Case School of Applied Science in Cleveland. Then he had a job I guess while going to school as a telephone lineman and installer. When he graduated he got a job teaching electricity at East Technical High School and also a job as a motion picture projectionist. For years he worked shifts. One day he would work the afternoon shift and the next day he would work the night shift, the evening shift. I don't remember his having been a high school teacher because I was born in 1928. He graduated from college in 1921 so he may have been through with that by the time I was maybe four or five. But he was a motion picture projectionist for a long time at least until 1945.
Then he started his own business with his brother of repairing and installing sound equipment in schools and industry and also motion picture projectors. He had a shop around the house. I guess he started his business after we moved from the home in Cleveland, Ohio in 1940 to University Heights, which is a close suburb. I went to the Cleveland Heights High School. Certainly at that time, and for some years more, he was a motion picture projectionist, but he also had this business, Gartech Sound Equipment. They built an addition at the back of the double garage of maybe about 25 feet by 20 feet and had a whole lot of equipment there.
Of course, by the time I was 12 years old, I was very much interested in science already. He had a lot of books around the house, engineering manuals and so on, which I read at an early age. I was a good student. My handwriting was terrible, but otherwise I was a good student. I learned to type at the age of seven or eight, because otherwise the teachers wouldn't accept my work. I was not very good at athletics. I liked to swim, but I couldn't run very fast. People would not choose me first off for their baseball team, but I liked science. I read a lot. I built a chemistry lab at home in the darkroom. I had hobbies of photography and whatnot. Then I graduated from high school in 1944 in the accelerated program because there was a war on. I graduated from Case in physics in 1947 for the same reason, but by then the war was over. I had a half- scholarship to Case I guess and a scholarship/fellowship at the University of Chicago for my graduate work.
What are your wartime memories?
Well, I don't remember very much. I knew it was a bad thing for people to have a war. A couple of my uncles went to war. Neither one of them was wounded, but they both saw combat. It was not a good experience. We're Jewish, so after the war it became very obvious that the Nazis had had a campaign to annihilate Jews and some other people as well, which made it even more horrible.
What was your reaction to the first atomic bombs and that announcement that happened in the summer of '45? Do you recall that particularly well?
I thought it was very interesting and a big achievement that one could obtain energy from the nucleus. Of course anybody who had been following it would've known that, but I hadn't. I guess in 1945 I was already in college, but people didn't talk about nuclear energy. People who knew anything about it realized it was classified and most of those people weren't at the universities anymore anyhow. They'd gone into war work of one kind or another. And, during the war, I entered Case in 1944 and several people were away doing acoustics, undersea warfare, or at the radiation laboratory at MIT. Exactly what they were doing was not clear. Some of them were maybe in the Manhattan Project, but I didn't know about that either. So then they soon came back and a couple people came back. I don't have a good memory for what was going on. It was clear that Enrico Fermi was a very good physicist so when he turned up at the University of Chicago I decided that would be a good place to go and that was right.
Is that the deciding factor that pointed you towards Chicago?
Oh yes. Chicago had an all-star cast. At the end of the war the Chancellor, Robert Hutchins at the University of Chicago used his endowment or other money to hire all kinds of people from Los Alamos. Fermi, Harold Urey, a lot of other people, Sam Allison. So he put together an outstanding Physics Department, and in fact some other departments too. Many of those people had been there during the war at the "Metallurgical Laboratory", which was part of the Manhattan Project. And of course Chicago is where we had the first man-made chain neutron-carried chain reaction under the West Stands in 1942. Fermi did not go to Los Alamos as soon as it was created in March or April 1943 because he was busy working from Chicago on the building of the production reactors at Hanford. He and the family settled in Los Alamos in August, 1944.
Before discussing Fermi, one general question. Why did you choose physics? I understand that your father had a scientific and engineering background and it was something you were interested in, but why physics and why not chemistry or botany or something like that?
Well, my father was an engineer and he wanted me to do engineering. But I figured if I were going to do engineering I'd need to work for somebody on some project or other, and I didn't want to do that. I wanted to do something on my own. You need to know a lot (or at least I thought you needed to know a lot) before you got into botany, which did not interest me, or chemistry or whatever. Whereas in physics it's amazing how much you can do with very little knowledge and so that hooked me. I had a professor of mathematics at Case who wanted me to go into mathematics, but that seemed to me to be too abstract. All these were just arbitrary decisions and, for most of them, they were right.
Tell me about arriving at Chicago and beginning work with Fermi.
That was a difficult era in the early post-war society. I'd just gotten married to my high school sweetheart, Lois Levy.
What year was that?
1947, April 1947. So there were two of us, and there was a big housing shortage in Chicago. We'd never had to look for a place to live. We had lived with our parents until then. I hunted for an apartment and there wasn't anything that was affordable because we had very little money. My wife got a job. She worked for Blue Cross. That's in the previous interview, Blue Cross and Blue Shield. Then we were back in Cleveland for the summer of 1948, I guess it was. My uncle was the chief accountant and leading controller for the Ohio Bell Telephone Company. He got Lois a job as, what do you call it? When you call on the telephone and you want to talk to somebody, so a telephone handling person, a customer relations person. Like all the other women, she was "Miss Allen" on the phone. For the first year in Chicago, we did not have a permanent place and we looked in the faculty club or elsewhere for people who were going away for a couple of weeks on vacation and we would sublet their place. Then we had a shared apartment. That didn't work out because the landlady was terrible. I think we lived in 13 places in 11 months, and some of them twice, because there's a hotel that we could go into.
I returned to Case in the summer of 1948 in order to work on two inventions I had made before going to Chicago. One of them was more than enough for the summer, in fact, I didn't get very far. The idea was to have an intense projection TV source. If I had that, the size of a frame of 35-mm film, obviously it would be possible to project this onto a screen of adequate size and brightness. In fact, it is not so obvious, because my light source was to be a phosphor, like that in an ordinary TV tube. But it would be illuminated by a much more powerful electron beam than is normal.
In order to have such a bright phosphor, one would need to select the material; one would need to have a transparent face plate which would conduct heat well enough so that the phosphor wouldn't overheat, and the beam would somehow need to be kept from spreading due to Coulomb interaction as is the case with a scanning pencil beam which is the basis of all normal TV tubes.
Thus I planned to use a fine-pitch metallic grid just in front of the phosphor screen. The wire of the metallic grid would be covered with insulating material, and I would write a charge pattern with a normal scanning beam. The charge pattern would then modulate spatially the flood of electrons from a large cathode, which would otherwise provide constant bright illumination from the phosphor.
Of course, single-crystal diamond (with a thermal conductivity better than copper) would be the ideal face plate, but my budget would not extend to that. So I used Z-cut quartz, which as a very substantial thermal conductivity. I still have those pieces of quartz.
I machined a brass "tube" and pursued some of the other efforts, but three months and one person is a factor 10 short of the effort that would be required to achieve this. So it languished.
My other invention was to grow diamonds at near room temperature and at atmospheric pressure. I wanted to start by showing that I could grow one of the unstable crystalline forms of sulphur in this way, by super-saturating a solution and presenting it with a seed of the unstable form.
For diamond, I had looked up in the literature the solubility of carbon in lead as a function of temperature and decided that I could saturate lead with carbon (graphite), super heat the lead so as to dissolve any tiny nuclei of crystallization, and then bring the molten lead in contact with a diamond seed. Unfortunately, when I tried this experiment after leaving Chicago and going to IBM in December 1952 (so I must have tried it in 1954), I discovered that the "standard" data on solubility of carbon in lead was wrong by at least a factor hundred. So I would have been driven to the use of molten iron, in ceramic tubes, and I gave that up as well.
So you were really just rotating around from place to place during that time.
Right. Then the nephew of a friend of my mother's from Cleveland and his wife and baby had an apartment, a basement apartment that had been converted from a laundry room in a real apartment building on the South Shore, 7715 South Shore Drive. He was a veteran, so they were moving into better quarters in the Quonset huts on the campus, so we got their apartment. That was real heaven.
We stayed there until our other friends, Harold Agnew and Beverly and their daughter Nancy he got his Ph.D. with Herb Anderson and Enrico Fermi. That's how I met the Agnews. They were going back to Los Alamos. The housing shortage was really so tight that he was walking around the neighborhood and looking at old houses there. He found one with a first and second floor veranda. He went to the owner, just approached him, and said that if the man, Mr. Beaudry, would buy the materials that he would enclose the space and the staircase and make an apartment out of it, and then we could have it when they left. When they were ready to leave, it was the fall of 1949 because our son was born soon after we moved in. They were ready to go back Los Alamos and he told Mr. Beaudry that I was reliable and handy too. We inherited that apartment, which was much more spacious.
Then eventually, when I got my degree, I moved into university housing right next to the Fermi, not the Fermi National Accelerator Laboratory, but the Institute for Nuclear Studies is what it was called. That was built in 1948 and '49, so when I got my degree in December of '49, I moved from Ryerson Hall on the campus across the street at Ellis Avenue or so to the Institute of Nuclear Studies where I had a small office and a lab of my own. I began to do other experiments. My thesis was the first measurement of beta-gamma angular correlation, of the gamma ray that follows the beta ray in nuclear decay. People have been doing gamma-gamma and angular correlation. Then I moved on to work with the new synchrocyclotron or the 100 MeV betatron before the cyclotron was ready there.
What attracted you to that particular area of physics, of the early years of high- energy physics?
Well, that was the frontier so when you're first in graduate school you learn physics and then, as I think I mentioned in the Aaserud interview, after I'd been there for a few months there was this big risk which exam should be taken in which order. I decided that I wasn't going to get ready for the qualifying exam, which I was sure I could pass. I would take the basic exam (that is the Ph.D. exam) first. They said, "Yes. We advise against it, and if you fail you're going to have to step back and take the qualifying exam and then the basic exam. So you would lose two years or whatever." Anyhow, it was a very trying time. Nevertheless, I didn't like just theory. [Interruption] Where were we?
We were talking about Chicago and I was about to ask you about—
Why I was in particle physics or whatever?
Well, it was a frontier that's what people were doing. Actually, learning courses there wasn't any lab work at graduate school. So I went to Fermi and said, "I'm taking your course. My fingers are getting itchy. I really like experimental work and if you're doing something I could help with, I would be glad to work for you." He took me on. He was doing some work with Leona Marshall, they'd done neutron work together during the war. She was in fact at the first chain reaction. It was great. They were looking at positron decay, positron and things like that. Their positron source was a cotton thread soaked with sodium-22 solution which they sealed into a Geiger counter so that they would have an initial time and whatnot. But of course, Geiger counters were so slow. You had microsecond time resolution with them and they blocked for a hundred microseconds. They were beaten in that experiment by Martin Deutsch at MIT who was using scintillation counters. Fermi, who knew all kinds of people, got photo multipliers for the scintillation counters. I grew crystals for them.
What kind of crystals?
Anthracene and stilbene and naphthalene crystals and whatnot. So we got to work, but Deutsch had some experimental photo tubes, end-window photo tubes from RCA. Fermi got those tubes. That experiment was interesting. I decided that really it would go a lot better if you had not microsecond coincidence circuits "Rossi Circuits", which had been invented by Bruno Rossi in Rome in the 1930s and had been used throughout the war. So I invented some with nanosecond resolving time. I published that and it became standard for 20 years in the nuclear physics and particle physics business.
Joining the RCA photo multipliers then with modified Rossi Circuits.
Well of course, the coincidence circuits take pulses. The photo multipliers convert light to electricity, and the pulses that you get from anthracene or whatever are maybe 10 nanoseconds long. The ones that you get from sodium iodide are considerably longer, 500 nanoseconds, but you can make standard pulses from them if you determine their shape. That was the field that I largely created together with pulse generators and systems for handling these very fast pulses. That was useful to everybody.
What was Fermi like to work with?
A very wonderful person. A very informal person who got up early in the morning. They were building the cyclotron somewhat later, 1948 or '49. The engineers were building cyclotrons, but you don't buy cyclotrons off the shelf. They bought a betatron from General Electric, but the Chicago team built the cyclotron. Leroy Schwartz was the engineer in charge of building the cyclotron. He was mostly mechanical although he had some knowledge of electrical things. It was a synchrocyclotron, because at 450 million electron volts it's almost half the rest mass of the proton. They whole idea of the cyclotron, as invented by E.O. Lawrence, was that you have a uniform magnetic fields. Our 450-MeV cyclotron had pole faces 15 feet in diameter, and you could crawl between them and we did. We'd start the particle at the center. There would be a D across, two Ds, which was just a metallic pillbox, a little pillbox. This would have an oscillating radio frequency voltage on it. But, as a particle moves in a magnetic field, there is a cyclotron frequency, which is proportional to the charge, the magnetic field, and inversely as the mass. A proton will have a frequency of about one and a half-kilohertz per gauss, or at 10,000 gauss would be about 15 megahertz or so.
So, if you have a 10,000 gauss magnet and you put 15 megahertz on here and the protons accelerated this way and goes around shielded from the electric field. By the time it comes around here, it's at that point then it was accelerated here, it's accelerated there, too, and so it goes around in a slightly bigger circle. They don't have to be separated orbits. Each time it gets an additional kick. That's very nice because it's automatic sort of so you can tune your cyclotron to accelerate protons or deuterons or whatever. But this mass changes relativistically, and in fact, at half the rest mass, the mass is 1.5 times the rest mass. So what started out to be 30 megahertz (or at most 20 kilo gauss,) is now down to 20 megahertz. So Edwin McMillan, and independently V. J. Vexler in the 1940s invented the synchrocyclotron where the frequency would start high and then gradually over a 60th of a second or so would get lower. As the particle went more slowly around as it got higher energy, it would still be in phase, in synchrony. That would also allow you to reduce the magnetic field as you went out, which provided more vertical focusing to keep the particles in the horizontal plane. Still a wonderful thing, but when you do this, whenever you want to have an oscillator that oscillates at one frequency you find out that it likes to oscillate at many other frequencies. That's particularly difficult if you have to sweep the desired frequency like this, because if there's a parasitic oscillation any place you shoot through it. It takes out all of the energy and you don't accelerate and it's very bad.
So there was a problem and the engineers would come in and tell Fermi about the problem in the morning. In a very Socratic method, he would make sure he understood it. By the time he understood it, they would understand it, and so he would work out for them the experiment they should do that day to see whether the approach that he proposed would work. He would write it up in his little book. Every page was numbered and there was an index to his notebooks.
His "memory", he called it. He died when he was 53 years old.
In 1954, yes.
1954, which was a terrible loss. He was as I say, very informal. I used to have lunch with him everyday. When I joined the faculty we'd go up to the faculty club. Even though I wasn't in the habit of having lunch, it was so interesting to have these conversations.
So, despite his position in the scientific community, he was approachable and he wasn't a distant sort of advisor in any sense then.
No. Of course when I was doing my work on beta-gamma angular correlation, which used the techniques that I had developed to help him in his experiment, the only new thing I needed to do in order to detect the beta rays was to have a very thin source so that these electrons would not be scattered. I needed to know their direction and I needed to detect them without having them go through much material, because a beta ray will only go through a couple of millimeters of much material. I cleaved little slices from stilbene crystals and glued them to cover the end-window of a photo multiplier. Then the gamma counter could move around this. It was a lot of fun. It didn't take very long to do the experiment. Then you get measurements at different angular points. Now you can analyze in terms of the Fourier components and cosines of angle, but you make measurements at various points so you need to Fourier transform the measurements.
So the particles are coming out the side and you would move the counter at various points.
Yes. Because here was the thin foil, colloidal foil that I made myself. What you do is to dissolve collodion in acetone and then have a drop on water and it spreads out on the surface. This had been done by many others a standard procedure. Then you have a wire frame, kind of a coat hanger that you've submerged first. You spread the film. You pick it up. Now you've got a double film hanging from the frame. Then you can adhere that to the aluminum ring and put the material on it, usually a stable material, a sodium or zinc, salt or whatever. Then I took it to the reactor at the Argonne Laboratory and they radiated and made my radioactive material.
My angular correlation apparatus was a plastic cylinder about 25 cm in diameter, about 20 cm high. It just had O-rings to keep the air out when I evacuated it. I glued on the housing for the photo multiplier so the photo multiplier would sit in here for detecting the beta rays. You would have an O-ring to keep air from leaking in around it and voltages and whatnot, pulses that come out to the coincident circuit that does the analyzing of these things. Then another counter on a radius arm. This had a big block of scintillating crystal in there viewed by another end-window phototube. So you would make a measurement at this angle, this angle, this angle. Then you would Fourier transform by multiplying by sine theta and cosine theta sub I times the number of counts you had at the I-th position. Since the cosines and sines are not a normal set, you could analyze in terms of cosine theta, cosine two theta, etc., cosign three theta.
But you know there are only a few of those present. In fact there were never any in any of the measurements I made. I told my friend, Ted Novey, that there was another element that I didn't have time to measure. He was a nuclear chemist. He cared about these things and I didn't care about these things. That was thulium so he measured thulium and that was the first non-zero beta-gamma angular correlation. He used all my techniques. He was at the Argonne National Laboratory. But now we had these numbers, so I did some calculations. We had of course Marchant calculators. We were lucky, we had electrically driven calculators. We'd push the buttons and put in a number and then we'd push another button and it would do the multiplication for you and even a division occasionally if you had enough time to wait while this thing chunked along. My wife would come in at night and help me put in the numbers and do these calculations. This was done in Fermi's lab, just a corner of the workbench. I helped introduce the fast Fourier transform in 1963, a great advance, but of course, by 1963, people had electronic computers, which we didn't have in 1949.
Would he be interacting with you while you were doing this study?
Oh sure. He told me the apparatus would collapse, this fiber tube was too thin. I told him it wouldn't collapse. I had calculated it. What he was missing was that it was a tube of finite length with stiffening ends. If it had been of infinite length, he was probably right. It would have collapsed. That's exactly how you make submarines and things like that. Fermi made his own equipment himself when he could. He would turn things on the lathe and use the power hacksaw and so on, but he did not have a good understanding of electronics. I remember he arranged to get a surplus microwave test set with a little klystron in it and was doing some experiments just to understand by the microwave analogy to optics how these things would work.
Fermi was a very smart man. He wanted to make a Schrödinger Equation solver. You have the Schrödinger Equation that controls the evolution of quantum systems. For instance in single electron atoms, you know the potential energy as a function of radius is just the coulomb on potential. Of course it can be more complicated, yet if you're only asking about a single electron wave function in another kind of potential well, and if you have angular momentum; this can be solved also. There's one over R potential here, but one over R squared potential from the centrifugal force essentially. That was all solved, but Fermi was interested in the shell structure in nuclei as you learn more about the energy levels. So the stationary state there and another one there, it's easy to calculate those for independent particles in a well, but when you have multiple nucleons in a nucleus and a more complicated potential, there wasn't anyway to do it. He had an idea, you could draw an arbitrary potential, you know, one that you drew out higher
So he said, "Let's make a mechanical analogy to the Schrödinger Equation." The mechanical analogue would be really a compass needle. Here's a vertical pivot. You have a compass needle on it. This is the north end, this is the south end. Then around it a solenoid where the magnetic field would be proportional to the current. Now you have the analogue to the Schrödinger Equation where the angle of the magnet with respect to the axis of the solenoid is going to be the position in the well. The restoring force is then the sine of the angle times the strength of the magnetic field for a fixed strength magnet. So he said, "Well, let's make one like that." But then in order to capture the information you have to have a mirror on it, writing on a moving film. I said, "No, let's not do that. I can make a better one." I'll take these operational amplifiers and with two of them so the voltage comes in here and voltage goes out here. Then there's a potentiometer that is, then there'll be a feedback from here to, I guess this goes here. So this is going to be the potential that is the energy, the potential energy V. Then you have a system in which the voltage at this point is the output voltage, which is the input voltage multiplied by the position of this pointer on the potentiometer. So you have a graph recording mechanism that would bold the chart on which you would sketch potential. Then you would follow it and have an automatic follower, even in those days. This would be the electrical analogue instead of the mechanical analogue.
It's interesting from what you just said Fermi not being as electrically inclined, his coming up with more of a mechanical system, and yours being much more of an electronic solution.
That's right. He liked that so I actually built it and published it. I think he was the only one who ever used it aside from me a little bit and eventually Clyde Hutchinson. Fermi was a wonderful person to work with. I worked with him at Los Alamos because in 1950 I was appointed to the faculty of the Physics Department when I got my degree in December '49.
Teaching or only research?
Both. I was an instructor.
What were you teaching?
I taught all kinds of courses. I taught some freshman courses. I would teach Fermi's courses too. I taught some instrumentation courses, of course.
Did you like teaching?
Yes, I guess I may have had things I liked better. It's hard to do a good job at teaching. You have to either teach from somebody's book, which isn't all that much fun, or prepare your own. Fermi set a very difficult model to follow because he had his own courses. Of course there's the nuclear physics notes by Orear, Rosenfeld and Schluter, which was the nuclear physics course that Fermi was giving them. This was a marvelous thing, but tremendous effort in doing such a thing. I taught some of Fermi's courses when he had to be away. A lot of people, my students, were much older than I was.
You finished your work at Chicago quickly?
I finished in two years, right.
A remarkably short time.
Yes. Then in the summer of 1950 I was on the faculty so I was getting paid nine months of the year, which is what they did, and my family ate twelve months of the year. Fermi said, "I go to Los Alamos in the summer as a consultant. If you're interested in that work. You have to get a clearance, but it can be very interesting." He arranged for me to be a consultant to Los Alamos. In fact, he and I shared an office the summer of 1950. I was working in the physics division.
Did your family go to Los Alamos?
Oh yes, otherwise I wouldn't have gone. That was really wonderful because we had a government-furnished apartments with government furniture. They would give you amenities, lend you linens and blankets and anything that you needed so you only need to bring your clothes. Dishes, government issue dishes, refrigerator. We had a good time.
Had you traveled out west before the summer of 1950?
No. Actually I had gone I think to California for a job interview, but I'd never lived in a place like Los Alamos, which is a magnificent environment with the mountains and the greenery and rain and the clear skies. The whole population of the town was about 15,000. The next town was Espanola, on the way to Santa Fe. We explored the neighborhood. We're not aggressive mountain climbers or even mountain walkers or anything, but we had picnics in the mountains. A very nice place and made many more friends. Of course we had some from Chicago, our friends Harold and Beverly Agnew were there. Harold Argo and his wife Mary, who had both gotten their Ph.D.s at Chicago. But I think had left in 1948 so I didn't know them very well from Chicago. Marshall Rosenbluth, with whom I was to work closely was there.
Was he one of Fermi's students as well?
I don't think he was a Fermi student. He may have been a Teller student. A very smart person.
I guess he was probably a few years older than you?
I think so. And Marvin Goldberger, Murph Goldberger, is still a good friend of mine.
I understand that you were interviewed yesterday about the hydrogen bomb.
I don't want to force you to redo it, but I do have a few questions that perhaps weren't asked. This is a general one. Leading up to the time that you first went to Los Alamos, were you following any of the discussions such as Truman's announcement to build the Super? Were these things that you were paying attention to closely or was it more in the background?
Do you remember when that was?
The fall of '49 [January 31, 1950]. I want to say November.
Well of course, at the University of Chicago there were people who were very much interested in such things. People who had founded the Bulletin of the Atomic Scientists and the Federation of American Scientists. Chief among them were Leo Szilard, who was in the Institute for— There was the Institute for Nuclear Studies, Institute for the Study of Metals, and then it was the Institute for Biology, but it wasn't exactly named that. That's where Leo Szilard found his home, together with his colleague Aaron Novick. Szilard was a very ingenious person. He and Novick invented the chemostat, to which you would grow a bacteria of bugs continuously. They would be fed nutrients and gradually material would be flushed off including particle colony maybe with an exchange time of the order of 30 minutes to an hour with the bacterial generation time of 20 minutes or so. They would then grow up so it was really turbid. And then you would introduce some poison or something or reduce one of their food ingredients and they would mutate. They were mutating all the time, but of course you would favor the mutation that could coexist, so they made bugs that could eat things that bugs don't usually eat. It was a very interesting approach.
John Simpson, of course, was one of the people who after the Manhattan Project was dedicated to doing something about controlling the peril of nuclear weapons. He took two years off to lobby Congress and speak publicly about this. He was there too. I knew these people. In fact John Simpson was on my Ph.D. thesis committee, but I was not close to him. At lunch there would be talk about this, but so much was classified at that time so that the people who were involved didn't talk much to the others who weren't involved. Fermi was on the Atomic Energy Commission's General Advisory Committee, as was Rabi from Columbia University and Oppenheimer typically. So yes, I saw this. It wasn't known whether Supers were possible or not.
When I went to Los Alamos in 1950, of course, then I found out. Now it's said that I went to Los Alamos because I had made a suggestion to Fermi about nuclear weapons. And he said, "Well, that's very interesting. If you really want to know more about it, what's been done in that area, then you really have to go to Los Alamos and find out." I don't remember that, but it's printed, so it's probably true.
When I did get to Los Alamos, I spent the first several weeks in the classified report library reading all of the progress reports, weekly progress reports from the various groups during the war and after the war. Some of the work was on the Super. There were only a few people working on the hydrogen bomb during the war, much to the dismay of Edward Teller, who never forgave Oppenheimer for doing the right thing, which was to concentrate on the fission bomb. Because obviously you would need a fission bomb somehow in order to get the temperature to begin the nuclear reaction. But Teller maintained even in the summer of 1942 at the famous Berkeley summer study that the problems of fission bombs were solved and now they should go onto something really interesting. He's always had that view that what's interesting is what's important, and that's not necessarily true. Oppenheimer saw that the fission bomb would be good enough.
Of course, when they actually got to Los Alamos, which was the site where they were to assemble the bombs, the uranium-235 was coming from Oak Ridge, the plutonium was going to come from Hanford. They got little bits of plutonium and discovered that plutonium made in reactors can not be used in the plutonium gun where you have separate lumps of plutonium metal that you push together because of the spontaneous neutron emission from plutonium- 240. There's another problem. The alpha particles from uranium-235 come off at a very low rate corresponding to the lifetime of 700 million years; those from plutonium correspond to a plutonium lifetime of 24,000 years. That's a factor of 30,000. So the alpha particles in uranium- 235 can be ignored. After all, the neutron was discovered by Chadwick by using alpha particles on beryllium or other light elements. And the plutonium sources are much more powerful than the sources that Chadwick had. You get lots of neutrons from alpha particles from light nuclei such as beryllium or oxygen. There was a hope initially, and Segré‚ made these measurements at Los Alamos, that this was not inherent that by metallurgically purifying plutonium you could use it in a nuclear weapon. That hope was dashed when they found plutonium-240 has a spontaneous fission lifetime and gives off neutrons, and even if they made perfectly pure plutonium Anyhow. Then they had to find another approach, and there was panic in the laboratory. Then they went to the implosion approach that Teller of course ascribes to Seth Weddermayer and Weddermayer received the Fermi award for that. But Bob Serber in his LA-1, Los Alamos primer on nuclear weapons, already shows that implosion assembly was brought from the summer study in 1942.
But anyhow, be that as it may, implosion works very well. It is better than a gun in the sense that you can use smaller amounts of material because it actually compresses the material. I learned all of that of course. When you're learning it and you know something about some other field and you go into a new field, then you apply everything that you know to these new problems. There are many, many ideas that come up. The first one really was this problem of pre-initiation by a neutron too early. Gun-type weapons are quickly sensitive to that, and so were early implosion weapons. I said to Fermi, "You know, when they have battlefield of nuclear weapons," which is what we had planned, "you'd better understand how long and how far a previous nuclear explosion has to be " How far away and how long ago, "in order not to spoil your next nuclear weapon." So he said, "Yes." I worked out new means by which one nuclear weapon could influence another one at great distances. It's a rather complicated thing.
Is this the fratricide idea?
It's fratricide. It's not only the prompt neutrons, but it's the delayed neutrons, which are very important for nuclear reactors, but had never anything to do with nuclear weapons. About seventh-tenths percent of the fission neutrons are delayed up to a couple of minutes because they come from highly excited radioactive fission products. That tells you how long ago because you're still getting neutrons. Not only are you still getting neutrons, but even the neutrons that were there long ago caused fissions in your nuclear weapon that has not yet exploded. Those fissions continue to give neutrons, which are decaying. Both the source has a long time scale and the inherent generation has a long time scale.
I'm just curious. Why did you start working on the fratricide question?
I was just thinking about the implications of what I knew. It's not that anybody was worried about it and I tried to figure it. Then the long range actually comes from a really different problem that I identified. That is as the neutrons go through the air, they're absorbed much better than high-energy gamma rays. So if a neutron would give a high-energy gamma ray, then the high-energy gamma ray would have to be high enough to be above the photo-fission threshold. It just turns out that in nitrogen in air, about one percent of the neutrons captured by nitrogen give a 10.8 MeV gamma ray, and that's above the photo-fission threshold. That has much greater range than the neutrons themselves.
So despite losing a factor hundred in the number of gamma rays and having a smaller probability of causing fission than the neutron would itself, this is the predominant influence at long range. It turns out that some people picked this up and urgently did an experiment at the next above-ground test in Nevada. Others were wanting to make a missile defense system, relied on pre-initiating of the enemy nuclear warheads as they were coming in. But I didn't know that nobody told me about it at the time. The other thing I did in Los Alamos was to recognize that the DD and DT cross sections that is the reaction cross sections on which a thermal nuclear weapon would depend had been measured, 10 years previously, and at higher energy than is of interest in a thermonuclear bomb.
I began to build an apparatus to measure these at lower energies. It was a very nice apparatus because not only would you count the product, you'd have deuterium on deuterium, which of course in a thermonuclear weapon come about simply because of the temperature. But here when you're measuring you have deuterium gas, for instance in a cell, the thin window, and a deuterium beam comes in the precisely defined energy. But in order to know what the energy is in the gas you have to know how much energy it lost in the window and in the gas. I'd have a cell with two windows so it went through one window of the gas and then another window; very few of them react. Then in order to measure their energy on an absolute scale is difficult. I really wanted to measure their energy loss and so I had the energy detector, which would repel the deuterons if it had too high a voltage, really was the same as the energy supply, voltage supply for the deuterons. Then there was just a small power supply of batteries between them that would span the energy loss range. I got far enough on that in three months, together with all the other things, that the laboratory decided that was a good thing to do, but they didn't want to wait for me to come back and do it the next summer. That's really a wonderful approach to have somebody pick up what you're doing and carry it out.
Who picked it up?
Fermi remembered a person from Los Alamos during the war Brit, Jim Tuck, James A. Tuck. He was back at Manchester, I think. So Fermi, through his connections, got Los Alamos to invite Tuck to come and head this group for measuring the energy cross-sections. Tuck, of course, didn't have clearance anymore, so he came to Chicago and worked on the cyclotron there while his clearance came through. So Tuck and Tang and the rest of the people did some things on the cyclotron. Then Tuck was there, this was published then in 1954. It was Phillips, Sawyer, Stovall, and Tuck I guess. It was a nice experiment.
Did Fermi ever discuss with you his feelings about working on the Super? Because from reading the histories his opinions seemed to oscillate between being in favor of it and being opposed to it.
No, he didn't. He was on the GAC, [the AEC's General Advisory Committee] and that was a policy matter. He felt that policy matters and technical matters were different. He would talk to me about any of the technical matters, but not the policy matters. He was very cautious that way. When I had a safe for secret material, I had to have my own safe because Fermi's safe had not only technical secret material, but also policy secret material.
Despite sharing an office, you would have separate repositories for the things you were working on.
Yes that's right. Even back in Chicago in those days we had a safe as well. Stan Ulam would come to our office every morning. He and Fermi would go over calculations they were doing on the burning of cylinders of deuterium. Fermi used a spreadsheet, you know, a real spreadsheet, an accounts spreadsheet. Are you a physicist, also?
My background is in physical sciences before doing history.
What kind of physical science?
Okay. He would convert all the higher order differential equations, actually partial differential equations into first order differential equations simply by defining auxiliary variables. The different rows were time and the columns were temperature and density and neutron fluence and whatnot as a function of radius. He and Stan Ulam would take some initial conditions and he would fill in the initial row in the table. Then Fermi took the next couple of steps by hand with his desk Marchant calculator or his slide rule or whatever. Then they would then call in the computer. The computer was Miriam Planck [later Caldwell], a summer student.
Back when women were computers.
Yes that's right. She would take it away and come back the next morning and show them the spreadsheets. They would graph selected results and decide what to do next. Whenever people calculated this, it really didn't look good. That was the problem. Edward Teller had been pushing this since 1943. The more you knew how to calculate, the less promising it became. That's how it was in 1950.
I worked on some things, I had ideas then about magnetic compression, all kinds of interesting things about the partition of energy in layered media. It turns out that behind the shock, you know, if a shock wave is propagating into air and a strong shock propagating into materials. So the velocity of the shock is higher than the velocity of sound because it's subsonic in the material behind, but that's of higher temperature and density. There are the Rankine-Hugoniot Equations. It turns out that behind a strong shock the kinetic energy is equal to the internal energy. That's very interesting because the properties of the medium are gross properties. If you have a medium, which is layers of heavy and light materials, you know, steel and hydrogen, uranium and deuterium or whatever, and you have a shock going through that you know it's going to have a lot of energy. But then behind the shock the internal energy and the kinetic energy have to be equal. The heavy material doesn't have any internal energy. It has only kinetic energy. All that kinetic energy as the things rattle together is occasionally given to the light material, to the material side. A number of us were calculating things like that.
The other thing I did in 1950 was to devise techniques for using stable isotopes to make precision measurements of neutron fluence that's the time-integrated neutron flux in nuclear explosions. By putting the stable isotopes in specific places and looking for the material in the fireball or in the debris you could determine what was happening at various places.
Where you were working in 1950, did you sense a certain sense of urgency of working on this? What was your sense of atmosphere?
You mean because the President had decided to work on the Super?
Also the Fuchs case was also in 1950.
I don't know that I paid any attention to Fuchs as well as other unpleasant things. I know Fuchs came over during the war with the British group. He had worked with Peierls, I guess also at Manchester. Peierls was an excellent theoretical physicist. It was just a total shock to him when Fuchs was revealed as a spy. Fuchs was not a very warm, friendly person, but everybody liked him. He was the well-known babysitter and also the most used because he was single and a good physicist. I didn't pay any attention to that.
I guess people were sort of resentful of the President for having given them this [Super] task that was impossible so far as they could see. They were doing everything they could. And of course since Edward Teller now having this approval, now the onus was on him to produce or to complain about the laboratory for not producing, which of course he did. Then when I came back in May, I guess, of 1951 Teller and Ulam had had their idea of radiation implosion. That was in March. There's a paper.
This was the joint paper that the two of them had had.
Hydrodynamic Lenses and Radiation Mirrors is its title. Now things were really quite different and everybody believed that this could be made to work, but everybody had his own/her own idea. May I see that?
This is the New York Times article. This is the transcript of the interview that Teller gave with Keyworth in 1979.
Where'd you get that?
From Thomas Cochran at the National Resources Defense Council.
Alright. So everybody believed that would work, but everybody had his or her own idea about how to go about it, what geometry, what fuel or whatever. Teller wanted an experiment, just as it said in the article, and he asked me to devise an experiment. In fact he said this in 1981 at this meeting in Erice (Proceedings are "International Seminar on the Worldwide Implications of a Nuclear War" (July 1982). I said to myself, "What would be a persuasive experiment?" I usually like to think not only of one solution, but two or three solutions to a problem because otherwise if you think of just one then it's likely to be awkward, it might be wrong, or whatever. But if you think of two or three then you may have different ways of analyzing them.
You have competition, kind of internal competition. So I said to myself, "The best experiment is really a demonstration. If you try to make this too small, it's not going to be persuasive. If you make it too big, it's going to cost a lot of money. Let's see what it would take to make a real thermonuclear weapon." Because that's what you want to know. After you know how to make the weapon then you would devise an experiment that would prove the principles of the weapon. I took everybody's ideas and went away. I didn't go away. I just worked on them for a week or two and made the design choices. I talked to people and they gave me their strong views.
What people were you talking with?
There were not only Teller, but Marshall Rosenbluth and Conrad Longmire who was actually doing calculations, Harris Mayer. Harris Mayer is at Los Alamos after many years of being in Los Angeles. Rosenbluth is in LaJolla, and Longmire in Santa Barbara. We talked about the various approaches and what was important. I began to understand these things. With the radiation implosion you have...
The radiation, even though you see sunlight all the time and it has energy, but it doesn't have pressure so far as you can tell; but both the energy density and the pressure, which are equivalent, go as the temperature to the fourth power. On the surface of the sun of course it's pretty hot, but it's only half a volt and now we're talking about multi kilovolt temperatures instead of one kilovolt. A factor 2000 (between half a volt and one kilovolt) to the fourth power is little more than 10 to the 13th. It's the enormous flux. If you calculate the pressure, it is far more than the modulus of any material so you can squeeze things to a small fraction of their initial volume. That was the idea.
Now there are details in there where Teller says he had a feeling that squeezing wouldn't help because not only does the energy generation weight per unit volume grow as the square of the density, but the energy loss rate per unit volume grows as the square of the density. Because there are equal numbers of ions that you like to have hot and electrons that you don't want to have hot. If the electrons are hot then they radiate to the radiation field, and the radiation field then if it gets to be quite fat then holds the electrons down in temperature because of the "inverse compton effect". That's where they were. They were stuck because you might as well work with deuterium at normal density because it didn't get any better with compression and low density and this would allow the radiation that is produced by the electron bremsstrahlung to escape. But if the cylinder is too small it doesn't work because it doesn't capture the neutrons and the alpha particle energy. If the cylinder is too big it doesn't let the radiation out. There's really no size of cylinder where this would work.
But with the other approach, if you squeeze it enough so that there's so much possible energy production from the thermal nuclear reaction that it can overcome the loss and can bring up the temperature of the radiation both as well. That was the idea so I just put it together. I have a primary. I have a secondary. I have a radiation case and what materials should these be and what thickness should these be. How do you put it together? The simplest fuel was liquid deuterium in principle. Since I was a low temperature physicist for my particle physics experiments I built liquid hydrogen and liquid deuterium dewars with windows on them so the beam could go through and whatnot and then I would count.
Working with things at lower temperature wasn't difficult for you?
No. It was no challenge at all. I'd done it. I'd even made my own deuterium. When I say make my own deuterium, not through nuclear reactions, but I made it from heavy water because it turned out that it was a lot cheaper to buy heavy water than it was to buy the equivalent amount of deuterium as gas for some reason. I reacted heavy water with iron and got the gas. It was duck soup to design this whole thing so that it would satisfy the nuclear properties and keep the primary bomb warm while the deuterium was cold and stuff like that. While I was at it, after I designed that one and it was more or less approved they had created a Theoretical Megaton Group, TMG, which met every week or so. It was chaired by Hans Bethe. Which would look at various proposals, and so this was taken to TMG and looked at. There were a lot of criticism, but it pretty much withstood the criticisms and that's how it was made.
Did you present it to the TMG yourself or was it—
I don't remember. I remember once when the discussion was pretty far along I was at the back of the room. There were about 25 people in a rather small room. We were trying to decide how thick the outer casing should be. I had argued that it should be half as thick as they wanted it and they said, "No. Let's be conservative." I said, "No, Hans. There's no way that you would need this because in order to affect the inside, the shock has to go to the outside and then has to come back. You only need half as thick as this design shows." But they went ahead and built it with the big thickness.
The conservative approach.
Yes, but it was impossibly conservative because there was no physical way in which there could be any effect. It's like they're moving things farther apart outside the light cone. I want to be sure that in your event the explosion's not going to affect me. If I'm a mile away it takes three microseconds for light to get there so I only need three microseconds of separation. It's like saying you have to be two more miles away if you have three microseconds, but kilometers actually, but that's impossible. It worked fine, but I suppose people thought I was pretty brash.
Not intentionally. At the same time after that one went into engineering development, I decided that I should design these same things to be carried by airplanes. Much to my surprise I read much later in the draft of Herb York's book, [The Advisors: Oppenheimer, Teller, and the Super Bomb] that the AEC had made five or six early Emergency Capability Weapons using liquid deuterium according to this design. I had trimmed their outer case. I had arranged that they could be supported horizontally and would withstand eight times the acceleration of gravity. That's not so easy to do because you have the force of gravity and also the need for low heat transfer so I used long stainless steel bolts set at angles. Because you have half-percent contraction when you take something down to absolute zero or liquid hydrogen temperature, things like that. If you don't do things like that your bolts will break or it will not be properly restrained. I worked also on diagnostic experiments.
What does that mean?
When you have an above-ground or an underground test, if it's an experiment you have to know what's happened. You can measure the yields in various ways, radiochemical, yield measurements, or various other ways. But here we wanted to know a lot more details. I mentioned that I had pioneered this stable isotope measurement technique. That was first used in the summer of 1951 in the test series there.
This was the George Test?
Yes, it was used in George, right. And there were other things that were going on. There were neutron experiments that were used in George. I consulted with the people, many of them from other laboratories, Herb York for instance was at Berkeley. They were doing experiments on George to see, that was a little bit thermonuclear burning, but it's never been revealed what the concept was or the geometry. There were people from the Naval Research Laboratory. Ernest Krause and Montgomery Johnson, very capable people so I had a lot of fun talking with them. They had pinhole neutron imaging PINEX.
Like a pinhole camera?
A pinhole camera not so far from the bomb and this would be a shield that would keep the neutrons from coming out except through this pinhole. Then we would have phosphor, these same fluorescent materials so neutrons coming through would then cause this to fluoresce and you take rapid framing pictures of them. So I worked on the framing pictures cameras with Berlin Brixner, who was the great expert at Los Alamos. There were gamma ray pictures, too. You would separate the gamma ray pictures from the neutron pictures by using absorbers that would preferentially absorb gamma rays like lead, for instance.
But still using the pinhole idea.
Right. To absorb the neutrons you would use light materials that would transmit the gamma rays, very clever things. In fact there was at Eniwetok, at the Mike shot, there was a two- mile plywood pipe 6 or 10 feet square filled with helium in order to allow the neutrons to come out. You could do time of flight of the neutrons too. Then you could measure their spectrum very accurately. All kinds of wonderful things, but here the idea was radiation and to measure the temperature of the radiation, to measure the pressures, to measure the properties of materials, the opacity of materials. Because your job is to contain this radiation for long enough and so you'd like to confirm your theoretical concepts of opaque that hold back this radiation.
Materials when they get hot are much less opaque than when they're cold because it's the electrons that do the job of absorbing and then re-radiating the electromagnetic radiation. Free electrons don't do much of a job at all. There you have only the Thomson cross-section. It's about sixth tenths of a gram, it's on the order of a gram per square centimeter for the mean free path. You'd have to have many, many grams per square centimeter to contain radiation and radiation is traveling at the velocity of light. But when the electrons are bound to nuclei, then the cross sections are much higher. They can be higher by a factor of 10,000 or 100,000. The problem is that at these temperatures the electrons get ionized off. It's a very complicated problem theoretically. You would like to measure it so we set up radiation-driven shock measurements with fast photography in order to see how the shock penetrates through materials and when the radiation front comes through, and when the radiation driven shock comes through. That was all done by photographing the surface of the bombs themselves. Teller had asked me what could be done in this fast photography. I worked out a whole bunch of things that could be done for these details.
I'm curious about your reaction to how historians have treated your role in the development of the hydrogen bomb? I mean Teller for example pretty much leaves out your contribution in his Better a Shield Than a Sword book. Richard Rhodes, all the different people, who've written about this.
I've read the Rhodes book. I don't think I've read the Teller book. [laughs] In fact I knew that Rhodes was writing that book. I had read his Making of the Atomic Bombs, I thought was a great book. There were people I talked to who had been involved who I knew a lot about it; and the people who were directly involved also said it was an admirable book. When I learned that he was interviewing people for his Dark Sun, I wrote him after he had interviewed Marshall Rosenbluth and some time had passed. I said, "You know, I was involved and maybe I have a different angle, view of these things as some other people." And he said, "Well," I don't know, maybe he was all finished with the manuscript, but he said he'd interviewed enough people so he thought he had it covered. He was perfectly polite. But Wechsler, I guess, who figures strongly in the Rhodes book on the cryogenics side— Actually there's one thing which is wrong in the— and I don't know whether I sent you my e-mail to Bill Broad.
I don't think so. [This is in Garwin's file]
Maybe it's in the letter that I sent you now, but here it says about the Los Alamos regulars "They were burnt out from too many rush efforts to build and test prototype nuclear arms." That isn't what I said. In fact, Bill Broad was kind enough to send me excerpts from his article before he published. I wrote him back an e-mail and said, "There's some things that could be said differently or better, but this one is wrong. I said it was the low temperature physicists at Los Alamos, the experts in liquid hydrogen that were burnt out. That's what they told me."
The cryogenic people.
Yes. Ed Hammel told me. I just couldn't believe it, that they hadn't done physics for two years because they were so involved with George and they had to go back and do physics. I argued with them a little bit, but I didn't have much power, any power, so I did it myself. It wasn't very difficult and they presumably could've done it too. I wrote Bill Broad in 2001, and I wrote the head of Los Alamos, John Browne. Well, John Browne said he was glad that I had written him because that makes it clearer, and that the people there had no idea I was involved in those things. Also people at Livermore. I happened to see some correspondence from a person I've known there for 20 or 30 years, and he was writing another Livermore person who happens to be on foreign assignment. He said, "We've dealt with Dick Garwin for a quarter century and we didn't know that he was involved in that." But you don't go to a meeting and say, "Well, here's what I did when I was young."
It's very difficult to do history, and my case is particularly difficult because I was there only in the summers. I didn't interact with a lot of people. I would go off, I'd think about things. I would write a report. I'd talk to Carson Mark or Edward Teller or whatever. I was much involved with my family. When there was something that had to be done, for instance in radiochemistry, I had some good friends from Chicago, Anthony Turkevich and Nate Sugarman. Sugarman died young. But Turkevich, and he introduced me to the people who do radiochemistry at Los Alamos Rod Spence at that time. I was interested in that and I talked with them and found out that they were doing something wrong, which is a very funny thing. We fixed that up and they were happy. But there's only so many hours to the day.
And it was a matter of another invention, the neutron source, I would talk to people and make a proposal to the group or whatever. They in fact went ahead and did it. I would meet some new people and I could go and talk to them again if I wanted to, but I had a lot of other things to do.
These people remember fleeting interactions maybe and then the program just goes ahead. I was never involved after the very beginning. The Mike fuel wasn't chosen there were still questions as late as January of 1952 whether they would use liquid deuterium fuel. In the summer of 1952 then I was working on some other things, but Marshall Rosenbluth was doing calculations on the SEAC. It's the computer that the Bureau of Standards had in Washington. It had mercury delay-line memory, I think it had 64 words, but it was better than what was available at Los Alamos. I went and I spent two weeks in Washington with Marshall.
Using this computing device?
Yes, well he really used it. I was kibitzing. We would use different configurations of thermonuclear secondaries and see what the yield calculation gave. I would estimate on the back of the envelope. I got to be pretty good from first principles, comparing them with what the calculation gave. That was fun. I don't remember everything I did in 1952, but certainly nothing so significant as Mike or these neutron sources.
Did you have a particular reaction to the Mike test?
No. I was at the University of Chicago when I heard.
One thing that historians will be debating for a while now, especially with the resurfacing of this Teller interview, is the relationship between Teller and Ulam. I was wondering what insight you could shed on that?
It's very difficult. I mean Teller apparently in his testament says that what he really holds against Ulam is that, after the George test, Ulam went back east and he said, "This proves that the Super won't work." My wife and I had lunch with Francois Ulam just a couple of weeks ago, the first of May or thereabouts we were in Los Alamos and she's an old friend of ours. I asked her about that. She said, "Well he was just being facetious." In fact I read there's a secret letter of February 23rd from Ulam to John von Neumann. John von Neumann was a frequent visitor to Los Alamos, and he was going to come in March of 1951. This was February 23rd, 1951. Ulam says to von Neumann, "I have some new ideas here, and people at the laboratory are interested. Teller also, but that probably shows that they won't work out because Teller is interested then, and is favorably disposed to it now." That was certainly facetious. It could perfectly well be that when it was reported to Teller what Ulam had said, it got sufficiently distorted so that he really believed that Ulam had lost faith and that this wasn't going to work. Now Ulam, like Teller, was a person who was always interested in something new and novel, but he didn't care to see them worked out. He just wanted to go on to his next idea, whatever it was, whereas Teller wanted to see that everything that he proposed was actually built.
In that way, Teller was a pain to work with because he would be working with people who would be actually doing designs and they would laboriously figure something that would've had a chance of working. Many of Teller's ideas do work. Then he would change his mind. He would say, "No, no. I don't want to do that anymore. I want to do something else, like working on fission weapons that brings them in, putting them aside before they're ever built in order to make thermonuclear weapons". Though the fission weapon is a requirement for the thermonuclear weapon. I don't want to extrapolate too far. But that was one of the problems.
The other problem is that Teller really apparently felt that he had been his own worst enemy with his theorem that compression wouldn't help. Exactly— Would you turn it off a moment? [discussion off record].
[Here, Garwin is reading from transcript of Teller interview with Keyworth] "I ended up that the next theorem be equally impossible at 1000 times liquid hydrogen density. Now this was obviously wrong and I don't remember when I realized it was wrong. Because it made a simple point that if you compress sufficiently it could tolerate equilibrium. I was not happy. This was a decisive point and of course I was not happy with this until I found out why the scaling argument is wrong formally wrong because it would mean established through absorption with light quanta and it doesn't scale. In fact that's not important."
The question really is, where was Teller at this time? Teller really was at the University of Chicago, as I was, although he was away much more. He maintains that it was before March 23. March 23, I don't think that Ulam had an idea about using radiation. That letter from Ulam to John von Neumann of February 23, in my opinion really should be declassified. I don't see anything classified in it. I'll see what I can do in that regard.
That would be interesting to add the record on this.
Yes. Here, Teller says to Ulam, [NOTE: here Garwin is reading again from the Teller-Keyworth interview transcript] this is sometime in February "Ulam came to my office and said, "Let's compress the material." I said, "Yes." "And have a nuclear explosion and put deuterium in there." I said, "Stan, the simplest thing, and it might work, but I think something better. You should not compress mechanically. You should compress by radiation." He wouldn't take it I guess. He didn't understand it or didn't accept it or whatever. So I said, "Look, I will put down both of these ideas into a paper and we both sign it." In that paper I explained how and for the first time I wrote it down that compression would help. That you could compress by shock or else you could compress by radiation and it was much better to compress by radiation than to compress by shock. I did not say whose idea was one, whose ideas was the other. We both signed it. Then I had explained all these things to Freddie D. Hoffman, who wrote a much more detailed description. I think that report, again Freddie and I signed."
Then later on I guess he says he shouldn't have signed that report because he should've let Freddie sign it. I think I've seen this in other materials, but I'm not sure. Then they looked at lithium isotopes and whatnot. When the George shot had been fired, Teller says "I would've proceeded and happily acknowledged that this invention is due to Ulam and me if nothing more had happened. When the George shot had been fired, Ulam went around and talked to everybody in Los Alamos who would listen that the George shot had proved that the hydrogen bomb could never work. He traveled to the east and carried on this message. Why he did that, I don't know. Ulam didn't have the idea. He didn't write the paper. When it came a class to the decision he declared that he did not believe in it. To me authorship is a question of responsibility. If you signed the paper, you should stand up for it." I think there's a real misunderstanding and it'll never be resolved, but I think it's a misunderstanding. Of course Teller really had a deep understanding of all these things. That March 9 paper is a very good paper an excellent paper full of insight. I think most of that is from Teller. Teller says he wrote the paper and he probably did.
When you arrived at Los Alamos in the summer of '51, did you meet with Teller and then sit and discuss these things with him before you proceeded?
Oh no. I mean I went around and renewed acquaintance with everybody. Teller, I'd seen during the year at Chicago.
Would you talk about research on the bomb?
No, he couldn't talk about classified materials at Chicago. One interesting thing would be to know when Teller was in residence and where he traveled in the late winter and spring of 1951. When I was in Los Alamos early May, 2001, I reviewed my five-page report of July 25, which had this famous sketch. The famous sketch was not available when I spoke at the 50th Anniversary of Los Alamos in 1993. I asked them to review the transcript or listen to my talk or whatever. I said, "I have looked for it and I haven't found it." But they've found it now. In fact they found a copy in, I don't know, Fermi's personnel file or something like that. I read it again, but I didn't read my notebook entries.
What is this that you're flipping through?
This is just the notebook from the last few months (2001) when I was in Los Alamos because I wanted to look at the date of that. I'm sorry, it was July 25th, 1951. My paper is called ADWD 283. It's title is Some Preliminary Indication of the Shape and Construction of the Device Based on the Ideas Prevailing in July of 1951. Anything else you would like to know about that era? In 1951, I also spent a month in Korea because the Air Force was creating a Tactical Air Command and various people who were influential with the Air Force like Charlie Lauritsen and Joe Mayer at the University of Chicago and Jerrold Zacharias and whatnot wanted to know what kind of research and development laboratories the Air Force would need. Joe Mayer and I were sent to Korea to find out during the Korean War.
When was Project Vista [this was study of tactical nuclear weapons done at Caltech in 1951-52] prepared? This would've been after I believe or maybe about the same time.
I'd say about the same time. I think that project Vista may well have been pushed toward Tactical Air Command and battlefield nuclear weapons, but I forget. I was not involved with those things.
The March 9, 1951 Teller-Ulam report. It's work done by Edward Teller and Stanislav Ulam and report written by Edward Teller and Dennis Stanislav Ulam, 20 pages. It's number is LAMS 1225, but I guess that's well known.
That one yes. I have just a few follow-up questions. First of all your work on the hydrogen bomb and then Teller's later from the Italian conference in '81 and afterwards, did his acknowledgment of your role in it give you political clout in your work that came after that in terms of enhancing your credibility?
No, probably not, because nobody knew. It had a very limited audience, a very limited influence. Edward Teller and I have not gotten along very well at various times, even as early as 1967 '68 over the first ABM deployment. Teller was a big fan of deploying the systems and I said it really doesn't make any sense, they're not going to work. He firmly believes that once you start down the road you will perfect it and find something that works. I say to him, I remember writing this, "there's a misconception. You're thinking about the Atomic Energy Commission where the laboratories did the whole thing after the AEC wanted them to do it or not. Whereas here we're talking about military programs and once you set down this road, you're going to get what you asked for, if at best." The same thing with Star Wars, where Teller has said various things including it doesn't really have to work. But at the same time he was advocating nuclear explosion pumped x-ray lasers.
I remember on the 40th Anniversary of the creation of Los Alamos, there was a panel discussion there. It was Teller and myself, Hans Bethe, and somebody else. That was just two weeks I guess after Reagan's Star Wars speech March 23, 1983. Teller of course was very hot on the Soviet threat, and for building this defense against it. In fact, if you look at the March 30, 1983, New York Times, op ed you'll find a solicited op ed by me titled "Reagan's Riskiness" and a parallel one by Edward Teller, "Reagan's Courage.". They never told me that they were going to publish something by Teller on the same page, but they did. I remember at this Los Alamos session I had a transparency and I showed here's the problem with building these x-ray lasers assuming you could make them work. I had worked on them too. But the Russians, is the long time it would take us to have this capability could readily go to ICBMs that didn't take 250 seconds, but took only 100 seconds to get up to speed.
By using different rockets?
Fast-burn boosters. If they did that then since you're proposing to have [drawing] that's the earth. Here's a Russian ICBM and that's where it normally burns out, but the 100 seconds burn out there at lower altitudes by a factor 100/250. And we're proposing to have these things as submarines or whatever. So when you get word that this missile is being launched, your x-ray laser pops up here. It had to reach previously that altitude in less than 250 seconds, but now it has to reach this altitude in 100 seconds.
More than twice as far, more than twice as fast.
Yes. The usual answer that you get, the knee-jerk answer, is, "if the Russians can make fast burn-boosters, we can make fast-burn interceptors". But that misses the point because this ICBM has to get up to seven kilometers per second. The interceptor must get to a particular point in less than half the time, well in fact a greater distance in half the time. Whatever speed the interceptor had and it was about seven kilometers per second now it needs to have twice that, 14 or maybe 20. The mass ratio, the ratio between the initial mass and the final mass of a rocket by the rocket equation is the exponential of the velocity gain divided by the exhaust velocity. The exhaust velocity is typically a three kilometers per second typically if you ignore structure about two and a half kilometers per second. To go to seven and a half is "e" to the third, which is about a factor of 20. But to go to 14 kilometers per second there's another factor 20. It's a factor 400, now in fact it's more than a factor of 20 and you have to go find it. Typically you would have in one case 1% payload, the other case you'd gave .01% payload. The ICBM would sacrifice 5% in payload because they'd need a slightly stronger structure; but for the interceptor here the fundamental aspect of rocket propulsion is involved. You need a higher speed. Inevitably you have to have the square of the mass ratio that you had before; you sacrifice a factor of a hundred or more in interceptor payload. They didn't realize that.
It's very pragmatic.
Very fundamental. In fact at least a year, maybe two years later George Chapline from Livermore was at a meeting here in New York at the New York Academy of Sciences. Then we were having some kind of luncheon at the World Trade Center. I was talking to him. He said, "You know, x-ray lasers would still do the job." I said, "Well, how about fast burn boosters?" He said, "Well, we'll just make fast burn interceptors." You want to scream. You have to explain to him again. It's like explaining to freshman. Why don't they understand these things? They don't calculate it themselves.
More recently I think also in Erice, there was a big controversy. I think it was 1983. 1983 I guess. And in fact both Flora Lewis, who used to be the Chief Foreign Correspondent for the New York Times, and Mary McGrory of the Washington Post were there and so was not only Edward Teller and some other people from Livermore, but Dixie Lee Ray who had been head of the Atomic Energy Commission or whatever it was called. I got into arguments. I explained why Star Wars wouldn't work and it would be against our interests though it was obviously after, it might have been the summer of 1983. There were Russians there. Dixie Lee Ray accused me of being a traitor because I opposed something that the government was wanting to do. Edward Teller told my wife he didn't "know what had happened to Dick. He's become a war monger because he's not interested in defense, only in offense."
That's a very strange perspective to put it that way.
Edward lets his political goals, pragmatic goals get in the way, I think, of analysis. Leo Szilard was at least as smart as Edward. In fact, I remember Edward at an evening party at Murph Goldbergers' in Chicago probably in 1952 or so. He said to me, "You know, the way I (Teller) use Freddie de Hoffman to calculate, that's what Leo Szilard does with me." So Leo Szilard used Teller for calculating because Szilard was even less a calculator than Teller. But Szilard, he certainly had ideas of things that he wanted to get done or things that he didn't want to get done, but he would never attack a persons character or positions or have whispering campaigns or whatever. Teller, in the past, has not been inhibited to do such things.
Let's just jump ahead to the future. After leaving Los Alamos you're back in New York at Columbia. You're working for IBM. You were doing some work with muon beams. My understanding is that that then is how you got involved with Lederman and the work on parity.
That's fine. I left Chicago and the high-energy physics, (particle physics) field because I didn't like the sociology. I didn't want to have to tell people six weeks in advance what I wanted to do and get it approved or not. I wanted to be able to work on the things I wanted to work on. I didn't like the idea also of applying for research money from the government because then you had to tell people even longer in advance. I guess I felt that if you had a big program supported by the government, then they ought there to give people more freedom. The problem was you had a shared resource.
In terms of beam time.
And experiments are interacting. Some of the experiments that we did required internal targets and different positions of the targets. Very interesting, but I didn't like that. I wanted to do it by myself.
Just to set that up, when did you begin to notice that sociology developing, the big teams and the long lead times?
These teams were typically four or five people, that was a big team. The long lead time was six weeks. I'm very impatient.
So nowhere near what it's like today.
Yes. I didn't go into physics in order to work with people and have other people's decisions influence my work. I decided in 1952 that low temperature physics had not changed since the war and that was a field where one could do some really interesting work so that's what I wanted to do. Even though they had some low temperature physics at Chicago, it's a lot easier to set up some place then to build a cyclotron. Through kind of an accident.
Emilio Segré‚ had been asked by IBM to start a Solid State Physics Laboratory at Columbia University where IBM had the IBM Watson and Scientific Computing Laboratory. They had brought in, after the war, electrical engineers from the MIT Radiation Laboratory and built some small computers and other things. Wallace Eckert, a very capable astronomer who introduced the punched card into scientific computing, had a small advisory group which included Polykarp Kusch and Rabi at Columbia. They said, "Solid State is where the future is." Wallace Eckert thought about it and accepted it.
I came to join that new laboratory, new incarnation when they were modifying a former fraternity house, 612 W. 115th Street. The laboratory had been at 612 W. 116th Street, but we needed more space. I came in December and unfortunately construction was behind schedule. I should have came in April, instead of December. When I left Chicago, I had several things going. I'd done some experiments with cyclotron and the betatron. I had built all of this apparatus that was standard then for 20 years in particle physics, multi-channel coincidence circuits, fast pulse systems, whatnot. I even had Chicago make one of my coincidence sets and IBM paid them $500, so I had it sitting around.
Of course I knew the people in the physics department at Columbia because I had lunch with them everyday and we would talk about physics. Leon Lederman, I guess I'd known slightly, but we became very good friends. I would go also not only to the faculty club lunches, but the whole physics department lunches at Chinese restaurants on Fridays. I gave a few talks at Columbia about my work. I had graduate students and was working on various aspects of low temperature physics. Then in the summer of 1956, I knew T.D. (Tsung Dao) Lee also and Frank (Chen Ning) Yang because we had been graduate students together at Chicago. I knew them very well.
That's an interesting connection.
Even more interesting connections because while we were still graduate students or maybe I was an instructor, they and I and two other people entered at one of the rebus contest that Disabled American Veterans put on. We were going to win $64,000 or $100,000. We actually bought a two-volume set of the dictionary and tore it apart. I was the specialist I think on five-letter words. Never mind, that's another story. But I knew these people extremely well.
Of course I would go to the seminars at Columbia, whatever they were. When Lee and Yang gave a seminar probably in August or whenever about non-conservation of parity and that they had studied the literature. Really there was no evidence for the conservation of parity in the weak interactions! In that paper they have two experiments proposed. One is a polarization of nuclei looking at asymmetric beta emission, and the other is the pi-mu-e. Pi-mu-e decay asymmetry. The electron asymmetry was expected to be a lot smaller, because if parity is violated only a little bit, it's violated in the pi-mu-e decay so the muons are only slightly polarized; and then if it's violated only a little bit in the pi-mu-e decay, then the new electron is only slightly asymmetric. If parity violation were 5% effect, then it would be a 5% squared or one part in 400 effect. And so that isn't what you want. People started in photographic emulsion tracking muons and seeing, but there're all kinds of scanning bias there. All kinds of people did those experiments. So Miss Wu (Chien-Shiung Wu), knowing I was in low temperature physics, asked me I think in September— I've written this up incidentally. It's Adventures in Experimental Physics. I can send it to you. [in Garwin's file at AIP] It's also at T.D. Lee's 60th Birthday.
I'll look that up.
Now how to detail the asymmetry? So Miss Wu came to me and she said, "I'd like to do this cobalt-60 experiment. Do you have a demagnetization cooling set-up?" I said, "Well, funny you should ask. I'm working on one that is to be much better than anybody else's, but it's going to go slowly because I've recently created a superconducting computer project at IBM. Not only at our laboratory, but also at Poughkeepsie. (This was in 1956.) And I've got 100 people working for me at various places and I had this responsibility. I can tell you how to do it and I can tell you how we would do it and I would be very happy to do so, but my advice is for you to go to where they have an operating demagnetization refrigerator. You should go to the Bureau of Standards." She did, but she would come back and talk about her work and her problems.
Then at Christmastime more or less, this would be better in the things that I've already written because I paid more attention and it was not so long after the fact. She came back and at a Chinese lunch she said, "We have these warming curves after we cool the cobalt 60 in a magnetic field and demagnetize. We do seem to have an asymmetry, depending whether we have the spins up or the spins down, and that's a clear evidence of parity violation. Now we've got to track down all kinds of dirt effects."
Just to get rid of is artifacts?
Yes, because it's just a monotonic curve, and there are many other things that could create that. She was a very careful worker. I guess I was at that lunch.
But then the Friday before the January 4th I was not there at lunch but Leon was. I was in Poughkeepsie and came back. I want to about to sit down to dinner and Leon called and said, "I've been thinking. We may be dealing with polarized muons whenever we deal with muons coming from a cyclotron. The way you get muons from a cyclotron is that [drawing] here's the pole. Here's the orbit of the high energy particles. And as I said, I had my coincidence system, but it was just sitting on a shelf. Then the Columbia University cyclotron I think was about 480- MeV. We had a copper target here. Protons hit the target and they collide with the nucleons in the copper. For instance, negative pions would come out this way, positive particles curve this way so these are negative pions, which are produced directly in nuclear collisions. But, along the way, the pions only have a lifetime of about 20 nanoseconds and they're going not quite at the velocity of light. They'd go maybe about 10 feet before they decayed on the average. In here you're getting some muons as well. The pi's go to muons; minus to pi-mu-minus plus a neutrino, pi-plus the same thing, mu plus a neutrino. [Garwin is writing equations here]
But in order to come out of the cyclotron, and for various reasons you wouldn't like to use negative muons, but if you're using positive muons then the positive muons are bent more strongly. You usually use the pions I guess that are coming backwards, and then they can be bent the same way. You get these muon beams coming out. The pions have shorter range so you put in carbon filters. Ever since the Chicago days, people studied these pions, and you would measure their range by putting in absorbers. This is what the coincidence counters are for.
You would have slabs of scintillation plastic, or I invented also liquid scintillation cells, and these would be photo multipliers in shielding to keep the magnetic field of the cyclotron from bothering the photo multipliers. In order to detect a particle, you would have at least two such because otherwise there's all kinds of background radiation. These two pulses in coincidence from that and from this there at the same time in coincidence as determined by the coincidence generator or analyzer. Then you know you have a particle coming in, and you can stop that particle in a graphite block or whatever, a copper block. Then there are all kinds of things you could do. You could measure the life time for the decay. Mu-plus goes to e-plus, plus two neutrinos. This is mono energetic because there are only two particles. This has a spectrum of energy. So you can measure the spectrum of the electron and determine what's going on here. You can measure the decay time because now if you have another counter telescope here, here's one and here's another one, the electrons are very higher energy like 60 MeV or so. At a later time you'll get, in these counters, coincidences, and now you have one coincidence system looking to define the incoming particle. You have another one to define the delayed event. You measure this number as a function of delay from here with a time delay analyzer or something like that. That's what was going on at many cyclotrons including Columbia. Marcel Weinrich was a student of Leon Lederman.
I noticed he was an author on the paper.
Right. He was doing this experiment with negative muons. Negative muons go to negative electrons plus two neutrinos. What you should get is an exponential curve your number of counts as a function of time after the muon comes in and you should have an exponential curve. Or if you're plotting the logarithm of N, then you should have the straight line.
Just as it's fewer than over time.
With the positives— Anyhow, there were problems. He just could not get straight lines, just could not. So his thesis was being delayed because he could not get good data. Leon had this idea. He said, "What we need now is a way to look at the angular distribution of the electrons coming from the decay of the muon." Why is the muon being polarized? Because if you have a pion in flight, it's going at that speed and now a muon which is recoils with four MeV of energy going forward. Pion no longer exists. There's a neutrino that scoots backwards and the muon goes forward. Muon has almost the same momentum as the pion, but if it goes backward even though the pion may be 100 MeV, the muon is only four MeV, it's the velocities that add. The velocity was 100, square root of 100 plus or minus the square root of four. That's 10, that's two. Now adding is 144 in squaring. Subtracting is 64. This is the energy, so these backward muons are easily absorbed. They have much lower energy. They'll never get to our target. These forward muons carry the fact that they were born forward so they would be polarized. They're separated from the ones that go backward. If you had a pion at rest then muons don't go anywhere, the whole population is not polarized, but the population that is separated according to range are polarized.
But now you need to stop these neurons in a well-defined place, and the idea would be to mount these counters, this counter telescope on a table and move it around and measure just as I did with my beta-gamma angular correlation. The number of coincidences is a function of angle.
I told Leon when he called me before dinner, "I'm going to finish dinner; then I'll meet after dinner at the cyclotron. We'll see what we can do." That was Friday night. I got there maybe around 8:00 o'clock or 9:00 o'clock. We looked around for a table with a swinging arm on it and couldn't find anything. But I'd been doing nuclear magnetic resonance measurements on liquid and solid He-3. With a magnetic resonance measurement what you're doing is to use a magnetic field to precess the spin. You've got a whole population of these little guys, may be a muon there, may be not a muon there. But it would start it at a particular direction whether it was there or not. Then it has a magnetic moment and a spin, and at a certain time later it will be in another direction. You could use a fixed counter, a counter in a fixed position, and then time would be reflected in angle because here's the population of muon spins and it'd be going around just as the actual protons go around, but this is just the spin. The same thing we use in magnetic resonance imaging.
If you really know what the muon magnetic moment is, which we had no idea at that time, but electrons precess at 2.8 megahertz per gauss. Muons, which are 207 times as heavy, if they were only heavy electrons, which is one possibility, would precess 200 times more slowly because it's eB over mC. So instead of 2.8 megahertz for the 800 kilohertz, it would be divided by 200. It would be 14 kilohertz per gauss.
A muon only lives two microseconds and so to get it to precess once in its lifetime, two microseconds, that's a 500 kilohertz would require 30 gauss, which is a very small magnetic field. You have hundreds of gauss coming from really these little permanent magnets. I said, "Look, you know, we can't swing the counters. Let's swing the muon spins." I always walk around and look in junk piles. There was a Lucite cylinder about that big in diameter, about that long. We went and got some wire and lathe. I wound wire around it. At Columbia the counting room is right down by the cyclotron, but the apparatus where you actually set up the telescopes, the beams come right down by the cyclotron. But on a hill few hundred feet away is where you have your counters connected by cables. So with some trepidation we took a couple of those signal cables and put current down from the power supply where we were so we could vary the magnetic field through here. Then we counted at a fixed delay. Just as you would expect to see as a function of time, if the muons weren't precessing, here you would see something like this maybe.
So a curve then.
When the flashlight was pointing towards you, you would have at that particular time higher yield. No, these are not the same muons. It's not even a muon for every cyclotron pulse, but an occasional muon. But they would all add up. The time zero is defined by the incoming counter telescope. Event time is defined by an electron in the detection telescope. We set this up. I knew all about this apparatus. It was all my-design, the apparatus. We first pushed Marcel Weinrich off his equipment. We set it up to do this and the cyclotron at Columbia is down for the weekend.
This is work that you're doing over the weekend.
No, this is work we were doing Friday night. They turned it off at 8:00 a.m. on Saturday. We had had a significant effect, but we had been going to higher fields so this was as a function of time. At a given position this as a function with time at a given position. We went to higher fields. If you do it as a function of B or H then this is a counting rate at a given position and at a given time slice.
This should be a straight line then.
It should be completely flat. Of course you don't have to trace out this whole curve. You just choose two points on it where you think (I'm just making some assumption about the muon magnetic moment), where you think that there would be a difference, and we did. We took two values of B. As we accumulated data we saw there was a considerable difference. Then it vanished. The last half hour or so we were getting data and there was no effect. That was sort of discouraging.
When they turned off the cyclotron, we went down to the place where the counters were and we found that the wire was all down here around its ankles because the coil form had expanded from the heat of the wire. Then the wire had stretched and then the coil form contracted and the wire fell down. We didn't glue it in place.
Again over the weekend we went to Columbia, we went to IBM. We looked for our ideal, a table where you could swing the counters. We didn't find it, but I decided we would do a better job. Then there was question of what target would you stop these things in. Then because of my experience with magnetic resonance I decided it would be best to stop them in a metal, but didn't want to have something, which was very dense and had high atomic number because that could scatter the electrons. We got a graphite slab. There was a graphite target in here. This time I decided I'd wind the solenoid right on the graphite, which it makes it easier. We had the shield from the external field of the cyclotron so we just got some iron and mu metal and we made a thing like that. The cyclotron field coming down would get sucked into this metal, shove it down there, and go out the bottom. We have the same thing here. You can see this apparatus at the Smithsonian. They may have it in storage, but they have it. This worked perfectly. We started again on Monday morning and by, I think Leon went home at 2:00 a.m. or something on Tuesday. I called him up at 6:00 a.m. We had the paper all written. We had a beautiful precession curve. I told him, "Stockholm calling."
What was his reaction to that?
Well he came in. He lived only three minutes form the laboratory. He came in and we had 22 standard deviations significance on this curve. There's really no doubt about it. It was a wonderful experiment.
Wu's results were coming about the same time. I think she and her team had finished them right before the Christmas break or around that time also.
No, they were still doing their work.
They were still doing theirs.
Right. Because they had other checks to make, they were very careful.
How did you work out the priority?
Obviously we weren't going to publish before they did. So we published it February 15 and they could too. We were actually ready with our paper before they were, but we would never have done the experiment had she not had her results. Besides, T.D. Lee would have killed us.
Then a year after that or that year, I can't recall exactly what year they got the Nobel Prize, but it was either that year or the following.
No, I think it was that year, 1957. And of course you can only have three people in a Nobel Prize. Lee and Yang were obviously the people who deserved it. They had done all of this wonderful work. They'd devised experiments, told people where to look, and so on. Wu deserved it too, but I don't know. Maybe the prize committee thought that there were too many if they wanted to give it to them and us. I don't know anything about that.
Was there any previous position to award the prize to the people who had come up with the theory or the idea of non-conservation of parity versus the experimentalists?
You couldn't ignore Lee, and Yang because nobody would have thought of this had it not been for them. It turns out the reason that Marcel Weinrich didn't get straight lines is that the fringing field of cyclotron was making his muons precess. The effect was there all along and also in cosmic rays, the muons that come in cosmic rays are polarized too and the same thing had been observed there. People measured wrong lifetimes and even electrons asymmetry. Lee and Yang obviously deserved the Nobel Prize, but Wu and Ambler, Wu actually was the spearhead, but they had the experimental apparatus. We were done by January 9. They had preliminary results. We had our definitive results. She deserves the priority. Lederman phoned Lee, so January 8th must be Tuesday.
Yes it was.
In fact, Lee and maybe Yang came up on Tuesday noon or whatever, to look into the experiment. We told them about it.
Did you show them the experimental setup that you had?
Right. I told you about the experimental setup and then I got my coincidence circuit out of storage, which was much more capable than the ones that they had built.
So I resigned as Head of the Superconducting Computer Project. I told the IBM people that I was working on parity non-conservation now. I worked on various things with Leon, and without Leon when he went to CERN in 1958-1959. Then I went to CERN and did precision g-2 ("gee minus two") measurement with Francis Farley and others. I really had planned to go to CERN the summer of 1959 just to read in the library because I had been so busy. I wanted to find out more of what was going on and refresh my skills and whatnot. But Leon had been talking with them about designing an experiment to store muons and do the high-precision g-2 minus two measurement, which we had done. We had done g-2 two at that very first experiment when we did other g-2. But to store the muons for six microseconds or so while they went around and around thousands of times was a wonderful thing. The group, which had been organized and had done some preliminary analysis, then pleaded with me to lead the group. I relented and I did that.
What was that experience like given your predisposition to working alone?
Aside from the fact that I really wanted to be in the library so I resented doing this, it was great because it was a small group, five or six people. It was Niro Zichech‚, Frances, Carl York from the University of Chicago who didn't last in the experiment. Val Telegdi initially, but he had his own ideas. And Theo Muller, who was an electronics person, and Hans Sens, a Dutch person.
I said, "I'll do it, but if I'm going to do it I'm going to be in charge and here's how we will work. We'll discuss everything, but I'll decide. There may be two ways to do something and if I judge one is better we'll do that. I assure you all suggestions will be welcome and analyzed, but I will decide what to do." They bought that and then we parceled out the work. Niro Zichech‚ was shimming and designing this whole magnet, an 80-ton magnet six meters long. Muons were injected here, and we had a magnetic field gradient, which was fairly low along the main part of the magnet so that the muons would precess only a little bit and orbits would move only a little bit per turn. In the beginning we had to have bigger gradients so that muons coming in would lose enough energy in a target, so the next orbit would be separated and then they would be squeezed to precess. Finally at the end you had to get the muons out to analyze them. Otherwise if you don't do this they're just in this magnet and they'll just go around the edge because particles hate weak fields. But if they come to the end and the gradient is big enough so that the orbits are like this then it doesn't feel that. It crosses the field before it knows that there's a weak field there; so you have to have a very big gradient at the end. Niro Zichech‚ did a wonderful job of doing this. We all discussed how to do it and he did it.
George Charpak, that's where I first met George Charpak. I wrote a book in 1997 with him in French, Feux Follets et Champigrois Nucléairas. But we've been friends ever since 1959. He was assigned the job, he was the person who had invented all kinds of new counting means, which we didn't use. We used scintillation counter telescopes. He was assigned the job of making the polarization analyzer with these telescopes of conventional types. But the muon when it came out would be trapped again in a carbon block. Sometimes you would have a field in one direction on it, sometimes in the other direction. According to this, if there were polarization then you would be counting up these two points and so it would be most sensitive to polarization.
What we were looking at was this polarization measurement as a function of time because as the neuron is stored in the magnetic field, the g=2 means that the spin is always in the same direction as the velocity. But if there's a departure of g from two, the gyromagnetic ratio from two then the spin moves and so you wouldn't get any variation of polarization with time if g were equal two. You're sensitive only to g minus two. This worked like a charm and Frances Farley was given the job of making the analysis program. Theo Muller did the electronics and Carl York was going to build some of the circuitry. Using coincidence circuits now with semi- conductors. That was a poor experience because I said to him at our meeting, "Have you actually done this before? Have you made these things and you're going to copy them? I don't want to do anything new here." He assured me he had. Then they got delivered from the electronics shop and they had problems. I was going off to India for a week in the summer or whenever it was, just toward the end of my year at CERN. I just took some design paper with me and built whole new coincidence circuits and fired Carl York from this job because he hadn't been straight with us.
You were at CERN as Ford Foundation Fellow.
Ford Foundation Fellow, right.
In '59 and '60.
Right. In fact our magnet was delivered and I got this magnet. I helped design the magnet. How do you align the magnet? You got the rigging crew to come in and they put it here and then oh, I don't like that kind of interaction. Here's the magnet, it's six meters long and two meters wide or something like that. It weighs 80-tons; at the bottom we put three one-meter diameter pads. If you just do the division, if you have 100 pounds per square inch of air in these air pads so we had a thin steel plate put on the floor where the magnet was. The rigging crew put the magnet down and we waved goodbye to them. Then that night we rigged up the air hoses. I was alone and going to position the magnet, and I turned on the valve. Sure enough the floor wasn't level, so here's this magnet accelerator on the loose, and it was about to squash me against the shielding wall. So I hold out my hand. You know, 80 tons times one part in a thousand is still more than 100 pounds. I turned off the air and knocked it dead in its tracks. It was great fun and everything worked extremely well. Those people for the most part have made very good careers.
I think now might be a good time to stop. That covered all the questions about parity. Next time I'd like to cover some more recent topics. Thank you.