W. Conyers Herring

Kojevnikov:

This is August 5th, the year 2000. I am in Palo Alto. My name is Alexei Kojevnikov and I'm doing an oral history interview with Conyers Herring from Stanford University. I suggest we start with you talking about your career with Lillian Hoddeson in an old interview. That was about the time that you were writing your thesis. And try to follow up with the story of your subsequent professional career in Princeton in particular. When you got your Ph.D.?

Herring:

Yes. I got my Ph.D. in '37. I think it was granted early in the year, but I stayed until the end of the academic year in the middle of 1937. Jobs were not too plentiful then, but it was considered quite advantageous if one could do it to get a National Research Council fellowship. I applied for one of those and was fortunate enough to get it. I had specified going to MIT and working under Professor Slater there because Professor Slater had recently spent a sabbatical period at Princeton. And although I didn't find that my mode of thinking and his were especially compatible, I did have a tremendous admiration for his ability to pick a problem, sit down, start calculating, work it out, write up a paper very quickly. And he, during his time at Princeton, produced several papers that were of some importance. So that's where I went.

Kojevnikov:

If I could interrupt a little bit. I know Wigner at that time and Princeton in general was the stronghold of the Band Theory, and that's where more or less it was one of the birthplaces.

Herring:

Yes.

Kojevnikov:

What was Slater's attitude towards this? Was he a band theorist at that time?

Herring:

Oh yes, very much. Very much even at that time and of course became even more so later on.

Kojevnikov:

If you could specify, how would the approach at MIT differed from Wigner's? What sort of problems, maybe they preferred different sorts of problems?

Herring:

Well I think Wigner was more interested in what I felt were really fundamentals. Things like the symmetry properties of wave functions, which played a very important role in my thesis, and the many body aspects of the problem. The need to, even though one started with a determinant of one electron wave functions, now one really would have to take electron interactions into account and consider correlation energy and things of that sort. Slater was more interested in the one body problem—how does one efficiently calculate Bloch functions in crystals—and practical methods of numerical calculation and so on?

Kojevnikov:

Wasn't this the same year when he and Schockley, I think, wrote a paper on exitons?

Herring:

I'm not sure about the paper on exitons or Schockley. He did his thesis of course under Slater, and it was on the band structure of block solvent, I think. I don't recall work on exitons.

Kojevnikov:

Was Slater interested in specifically in what kind of work you were doing during that year? Or was he involved in it at all?

Herring:

He was not involved in it. I think he was interested and the results that came out, but I never talked with him in detail about the problems of my work or anything.

Kojevnikov:

Was he a team worker?

Herring:

Depends on the rest of the team, I guess. Yes, he was very good at organizing a group of people to work on problems of this sort and fashion and did so. He organized his Solid State and Molecular Theory group, as I think they called it. Over a number of years, they produced quite a lot of work. Orientation of all of that work being, I think, dominated by Slater's interests. At the time that I was at MIT, Slater had been working on a calculation of band theory using what was then called the cellular method, where one integrated the wave equation in a single unit cell assuming the potential for the electrons to be spherically symmetrical inside the central cell, and then applying suitable boundary conditions to join that wave function on continuously to wave functions in neighboring cells.

Kojevnikov:

Is it the same as Wigner-Seitz method?

Herring:

It was attempted as a generalization of the Wigner-Seitz method, but it had very serious defects as Schockley himself showed. The Wigner-Seitz method was basically a method of determining the lowest energy band state with spherical symmetry in each unit's cell. One did that by requiring a boundary condition of zero radio derivative of the wave function at the boundary of the sphere, whose volume equals the volume of a unit cell. It can easily be shown by a variational method that that gives a tremendously accurate value for the ground state energy in question.

Then in the Wigner-Seitz calculations, what they did was to try to use perturbation theory to calculate the amount by which the energy increased when you increased the wave vector from zero up to a small value. It kept the effective mass, in other words. That worked fairly well for monovalent metals where you had a Fermi surface that did not reach the boundaries of the unit cell of the Brillonin zone. But for multivalent metals, one needed to have accurate energies at the boundaries of the Brillonin zone and the next Brillonin zone. For those, one needed something more accurate than just treating the finiteness of the wave vector as a small perturbation because the wave vector was large. So Slater used this cellular method where he imposed boundary conditions of fitting of the wave function at certain isolated points on the boundary of the unit itself.

That produced a wave function that was actually discontinuous over most of the boundary of the cell. As Schockley showed very soon, that gave energies that developed rather large errors when the wave vector was large. So Slater realized that, the incorrectness of that and tried to work out something that would be better. He developed what he called the Augmented Plane Wave Method, which I won't undertake to describe for these notes here. It's described in all of the books. It again assumed a potential that is spherically symmetrical in each cell, but solved the boundary problem much better. He had his student working on making a practical calculation of energy bands, including the field 3D bands in metal copper. That was Marvin Chotero [?] who did a thesis on that and that was the first practical calculation of that type. Then Slater went on, I guess, to do or encourage others to do further calculations with the Augmented Plane Wave Method.

Kojevnikov:

Did he encourage you to do anything specific?

Herring:

He didn't try to talk me into it, no. I was working at first on the problem of trying to see if I could find out the perturbation and the energy that would be caused by an arbitrary infinitezimal[?] deformation of the lattice deforming it into essentially a frozen in phonon of given wavelength and wave vector. I was having all sorts of difficulties because when I moved the atoms away from their symmetrical positions, it was hard to describe the change of the wave functions near the nuclei of the atoms in terms of small perterbations. Because the displacement of the atoms could be even though small compared with the original spacing of the atoms might be sizable as compared with the distance from the nucleus to the first node of valence electron wave function. That's what led me eventually to try a method of making the valance selectron wave functions, the conduction wave functions orthogonal to the olecore [?] states of the atoms.

That led to the discovery that in the undistorted lattice, if you simply took a wave function say equal to a constant everywhere, made that orthogonal to all the course dates of the atoms, you got a pretty good approximation to the Wigner-Seitz ground state wave function. That was the basis of the orthogonalized plane wave method. When I realized that could be a convenient and even practical way of calculating band energies, I got a very good opportunity to apply it to a practical case because a very good friend of mine, Albert Gordon Hill, who was an experimentalist who had joined the department that same year as an instructor had an unfinished job that he had undertaken in theoretical work at Rochester working under Seitz of calculating the wave functions and energies for the metal beryllium, which has a valence of two.

We're finding difficulties. I told him I thought we could solve those difficulties with the orthogonalized plane wave method. We pitched in, both of us together, to make a calculation of that sort and published a very long paper eventually— actually, it appeared in print after I had already finished my two years at MIT as a National Research Fellow.

Kojevnikov:

Was it initially a two year fellowship or was it a one year fellowship?

Herring:

It was a two year fellowship normally, yes.

Kojevnikov:

At that time, what did you think? How did you see your future career? What were the perspectives available?

Herring:

Well, I hoped to get a faculty position at a university that had a fair amount of research, particularly in solid state physics. Of course this was difficult because there were very few jobs of any sort available and very few universities has a program in solid state physics.

Kojevnikov:

How would you describe the field at that time where solid state physics was pursued?

Herring:

Well, it was not completely unknown. In fact, as a graduate student I had seen some universities, particularly Caltech where I would spend my first graduate year. They did have a little program in solid state physics which the title sounded sort of funny to me, but it seemed to be a field that a number of people were interested in. Of course, when I became a solid state physicists myself, I took a great interest in it. But many universities did not have any work at al in solid state physics. Princeton would not have, yes it had a little experimental work and the theoretical work under Wigner. But that was all. In later years, a little bit later, I think they had no solid state work at all for quite a number of years.

Kojevnikov:

Did you regard yourself at that time all the way a solid state physicist? Or were you still open to the possibility to doing other kinds of physics yourself?

Herring:

I would value myself pretty much as a solid state physicist. Not everybody was that way. John Bardeen, who preceded me a little, still maintained a strong interest in nuclear physics when he went to Harvard. Although he published much more in solid state physics than he did in nuclear, he continued that interest. I did not ever publish anything in nuclear.

Kojevnikov:

What were the possibilities for finding a job at the universities? We're talking about the year 1939, 1940?

Herring:

Yes. I left MIT then in 1939 and did not succeed in finding a university position for a long time. Finally Princeton came through. They realized that I was still looking for a job and they offered me a one year job to hold me over so that I could look for a little longer, a one year job as instructor at Princeton. After I had accepted that, I did get an offer of a similarly a one year substitute job at a small college in Pennsylvania. I forget now the name of it, where a friend of mine was on sabbatical for a leave and I could take his place for a year. But I of course did not accept that because I had the opportunity at Princeton.

Kojevnikov:

Were you already married at that time?

Herring:

No. No, I did not get married for some years yet.

Kojevnikov:

So this was the year 1940, '41?

Herring:

Yes 1939, '40 at Princeton. Then I had to find another job. I interviewed at several places where one of the important things that I always asked about was what they had available in the way of calculating machines because I wanted to continue calculations of band structure in one sort or another.

Kojevnikov:

What kind of machines you needed at that time?

Herring:

There were three manufacturers that made good calculating machines. Monroe and Marchan, and I think there was a third, but I don't remember it. I used mostly Marchan's at MIT and I remember some places, one place that I interviewed for a job had only a machine that you had to turn with a crank, but the Marchan's and the Monroe's that I was more used to using were electrical.

Kojevnikov:

Were they physics departments or were they mathematical departments?

Herring:

Physics departments.

Kojevnikov:

Did you need assistance to do the calculating job, or were you doing only the numerical— ?

Herring:

I had to do all of it myself. When I was collaborating with Hill on beryllium, he volunteered to do all the calculation himself, but after that I was on my own.

Kojevnikov:

Just to give an idea, how much time would it take? Just give some typical or unit of calculating and how much effort and how much work would that require?

Herring:

Yes. It's a little hard to remember exactly how much time. But for example, to determine a wave function by the Wigner-Seitz methods, which is one of the simpler problems, one would choose a potential and tabulate it. One had a table of the potential, and one would have to put the potential in for each step of the integration. Then one had a program for integrating partial differential equation, which I used the— I think it's called the Neumoff [?] method, which was taught to me by John Bardeen when he was at Harvard and I was at MIT.

I was very glad to learn it because it gave surprisingly high accuracy for relatively widely spaced steps. So each step that I took required putting in the potential calculating a coefficient that involved the magnitude of the step and the magnitude of the potential more buying that into a value of the function being calculated. Then doing some multiplications and subtractions to get another step and then extracting the new value of the wave function from the result of that calculation. Then one had to do that over for a hundred times in succession, something like that. So it would take a matter of hours to do, not too many hours perhaps, but an hour or so to do one integration. And then having integrated it a number of times then to plot the values on calculate the value of the derivative at the boundary of the cell, and interpolate between the values I had found to get what energy would make it zero. Then I would have a ground state wave function for that potential.

Kojevnikov:

Calculation mathematics is a field of itself. Did you have to learn it from somewhere?

Herring:

Yes, there were books on how you do these calculation.

Kojevnikov:

Did you have to learn this in grad school or at work? At what approximate time of your career you first started doing numerical?

Herring:

I suppose about the time I was at MIT and had to do it.

Kojevnikov:

What materials or mechanical substances were you particularly interested and what determined the choice?

Herring:

Well, it's hard to say. I was of course interested in beryllium because it was available. Some of the ground work had been done. He had done found the ground state by the Wigner-Seitz method already. I would've been ultimately interested in other polyvalent metals. I was also interested in lithium. Although it's a monovalent metal, it's one for which the corrections to the Wigner-Seitz method were appreciable. I was interested in how great those corrections were. I was also especially interested in studying deformed crystals, calculating the elastic constants and ultimately perhaps the phonon spectrum. I never got around to doing a good job on the phonon spectrum. I did make some progress with my friend Hillard Huntington who had been a graduate student with me at Princeton and who showed up briefly at MIT toward the end of my stay interested in working on some of these problems. I think we tried to work on elastic constant problem for alkaline metals, but we didn't get as far as having anything publishable.

Kojevnikov:

Were you interested only in academic jobs or did you also look for industrial research at that time?

Herring:

I was mainly interested in academic jobs although word had gotten around that a few industries had a very good research environment, particularly General Electric. But I guess there were not many opportunities there so I didn't ever interview for any as I recall.

Kojevnikov:

To put it another way, did you feel that industry has any interest in the research that you were doing? Or was it at that time of pure academic interests?

Herring:

I felt that there was some interest in industry, yes, but I think I realized probably even as early as that that only a very large industry would be able to find it profitable to sponsor a program that had really basic research. So as I said, I didn't interview for any industrial jobs. When I got through with Princeton in 1940, I was looking at jobs at several institutions. One I interviewed for was at the University of Missouri. I think I also interviewed at the University of Minnesota and one or two other places. The Missouri job looked best. They were interested in starting to build up solid state work. They did not have terribly much there, but the leader of the department, I don't know if he was necessarily head at that time, but the leader was Newell Gingrage [?], who was a x-ray diffraction man, and fairly good. They promised me lots of support. They would get me a brand new Marchan calculator of just the right type and let me teach a graduate course in solid state physics and hope to develop further in the future. That's where I went in 1941.

Kojevnikov:

Was there at all the sense that the war was going on?

Herring:

Yes. The war was going on and although the United States was not in the war yet a lot of people started being hired by the MIT radiation laboratories to do radar work and so on and other jobs too. By the end of 1940-41 academic year my colleague at Missouri who had been hired the same time I had, Fidel Corson, an experimental nuclear physicist, had been lured away to work at the radiation lab. Shortly thereafter I got an invitation to work on anti-submarine warfare in New York City and felt that it was my patriotic duty to accept. So I had only that one year with Missouri.

Kojevnikov:

Did you have to quit your job at Missouri?

Herring:

Oh no, I was on leave—indefinite leave. So then I went to New York and that was quite a hiatus and I can tell funny stories about it. I was doing war work from the summer of 1941 until 1945. The pressure was so great during most of that time for all of the work on anti- submarine warfare and ultimately subsurface warfare that was including pro-submarine as well as anti-submarine that I had no time to think about much of anything else.

Kojevnikov:

Why do you think they contacted you abut the anti-submarine warfare?

Herring:

I guess they just went to the graduate schools and asked who has graduated with a degree in physics who you think is really very competent.

Kojevnikov:

So no particular experience?

Herring:

And I got a very good recommendation from Princeton and so they contacted me right away.

Kojevnikov:

What was the name of the institution or what kind of...

Herring:

Yes. Well, there was an organization of scientists for the National Defense Research Committee, NDRC, which covered all of the military areas. Most of their work was done by contracts with universities or perhaps industrial institutions to do this, that, or the other type of war research. I worked for Columbia University Division of War Research in an office that was staffed by officials under the NDRC Division 6, which was the division having to do with undersea warfare.

Kojevnikov:

Who were other physicists there?

Herring:

The physicist in charge of the whole office where I was (not a terribly large place) was John Tate of the University of Minnesota, who was the perpetual perennial editor of the physical review. Then they another, I guess the number two man was— I get a block on memory. From Bell Laboratories, Bell Telephone Laboratories. Then there were other 30 high ranking scientists.

Kojevnikov:

How big the unit was?

Herring:

I don't know in total. The office I was in occupied, it was at first downtown AT&T building, but later got larger and moved to the 64th floor of the Empire State Building and we occupied the whole 64th floor. I don't know, there might have been a dozen scientists and a number of secretaries and people like that.

Kojevnikov:

Was it mostly experimental work?

Herring:

Well no, not in a location like that. They have no experimental facilities. This was all theoretical or administrative.

Kojevnikov:

Where did you do your experiments?

Herring:

I never did experiments.

Kojevnikov:

Did you ever go to a submarine?

Herring:

I never was on a submarine, but I was on an airplane flying over the coastal waters to see how they detected submarines with radar and things like that.

Kojevnikov:

Did you say that part of the job was operation research?

Herring:

Yes. I got into the operational research actually spending most of my time on that. In the very early stages I got into a little bit of physics, but that was physics of underwater explosions. That was fascinating, but I didn't have an opportunity to do much more in it. Since underwater explosions belonged more in Division 2 of NDRC than in Division 6, Division 2 was I think explosives, I was made a consultant or something like that. I had some sort of legal attachment to Division 2 as well as to Division 6. Actually after working for a while there in New York just in the NDRC office with the Columbia University group, I was made also a part- time member of the operations research group, which was entirely different. Set up under Professor Phil Morris of MIT, an acoustics man, whom I had known from MIT of course, to try to do with headquarters in the Navy Building in Washington, D.C. So I did an enormous amount of train travel back and forth between Washington and New York. What Morris was trying to do was to do for the United States something of what the physicist Blackett had done in England.

Kojevnikov:

Was there much communication between the British and the American groups?

Herring:

Quite a bit, yes. We had visiting Britishers sometimes working with us and so on. Morris's group started out as an anti-submarine warfare operations research group. I remember about the first meeting that I had with his group, for which he chose Bill Shockley as the head of the effort, Morris explained to us that we were losing the war of the Atlantic and losing it fast. We really had to do something to cut down these losses to submarines.

One of the most important things to do was to combat the submarines operating in the coastal waters off the U.S. east coast. One of the most important weapons that we had and we had it starting at about the time that I joined the group was 10 centimeter radar. I don't know if you know the story of 10 centimeter radar and submarines, but the first radar that was practically used in naval operations was 30 centimeter radar. It proved very successful in locating submarines when they surfaced at night to charge their batteries.

But the Germans soon caught on to that and built receivers that would detect the radar signals when the planes were much too far off to detect the echo so they had plenty of warning to resubmerge. So the 30 centimeter radar was relatively useless. But by maintaining extremely tight secrecy we developed an operating 10 centimeter radar at MIT at the radiation lab that was very efficient. It was so sensitive it could even pick up from a few miles way or relatively short distance, a submarine periscope much less the submarine itself.

Then we succeeded in making a real killing of submarines in the Bay of Biscay. And the Germans asked their scientists, well, what could be guiding these allied planes that are killing our submarines? The scientist told them well radar that's 30 centimeters we've got that taken care of. It must be something else like infrared detectors or infrared suppressing paints and things like that. For a long time and for a whole year I understand after the German Army had captured Allied 10 centimeter radar in North Africa, the German Navy didn't know about 10 centimeter radar. That's because they had such tight internal security that the Army didn't speak to the Navy. That came out after the war. That was one of the things, but as I say we were mainly interested in operational research and I did operational research for the operations research group feeding directly into the Navy in Washington. Then I did other types of operational research.

Kojevnikov:

Was that with Shockley?

Herring:

Yes. Other types of operational research having to do with new underwater sound devices or new weapons devices or something like that having to do with Division 6 of NDRC. So I was working for two bosses on the same type of work really.

Kojevnikov:

What kind of problems was there in operations research, if you could specify at all?

Herring:

It was marvelous. Morris and Shockley hired people with brains and with the scientists type of attitude toward problems wherever they could find them. We had a great many, not only a great many physicists and chemists and so on, but a great many actuaries who were very good with the mathematics and probability theory. Of course everything had to do with probability theory. We had one chess grand master and all sorts of people. The group enlarged by the way too. Along side of anti-submarine warfare, they had a whole group a different person doing airplane warfare and another group under another person doing pro-submarine warfare. It was very active, very lively group with lots of ideas. In fact one of the most successful things they did was a ridiculously simply application of common sense that was pointed out and the argument, the salesmanship for it was made by a distinguished plant physiologist. Namely this is a man who became head of University of California Santa Cruz. Again I'm missing the name.

Kojevnikov:

We can edit that back in later to just get the names.

Herring:

He pointed out that ships went across in convoys, a large number of ships with a fleet of destroyers protecting them and moved across the Atlantic. The ships occupied so much space that it was much longer than the distance that a torpedo could travel. So the submarines would attack these convoys from the outside sinking the ships that were within the torpedo range of the outer boundary of the convoy. Of course by the symbols surface to volume ratio or perimeter to area in this two dimensional case, the larger you made the convoy the smaller the percentage of ships that would get sunk. Now of course making a convoy larger has a cost because it takes longer to assemble the convoy and so on, but you're saving a lot of ships that way. He did the statistics, convinced the Navy that that was right, they did it, and it worked.

Kojevnikov:

What were your problems there or what were your tasks basically?

Herring:

I don't know. I did a lot of things. I did some lab probability studies on effectiveness of different types of ordinances against submarines. The old idea of depth charges which sank very slowly which would drop from the rear of the destroyer. They had low yield because the submarine's here and the destroyer's here, they lose contact with the submarine after about a couple of hundred yards ahead. Depending upon the depth of the submarine because the submarine passes under the beam. So they have to decide where they drop the depth charges by extrapolating where they think the submarine is going to be by the time the destroyer gets up here with it's stern over the submarine. Actually, not even that. They have to decide where the submarine is going to be by the time the depth charge has sunk to the depth of the submarine. It's a long extrapolation and the yield is very small. So they developed systems of taking a much smaller charge, one that would have to actually make contact with the hull of the submarine in order to explode. But throwing in the head of the destroyer rather than behind and the things would sink faster into the water. Those did give a better yield, and of course we were calculating and trying to get a quantitative estimate of how much better they would do before they actually instituted the policy of doing it. And other types of ordinance that I won't attempt to describe. Also I remember writing a little report that got a very favorable reception just on the evaluation of the characteristics, the virtues and defects of some Japanese sonar equipment that had been captured and so on.

Kojevnikov:

Was it your first experience doing something different from basic research?

Herring:

Yes.

Kojevnikov:

How was the feeling, or how did you at that time ??? ??? ??? a difference?

Herring:

I was favorably impressed how much one could do on practical problems that seemed perhaps at first site rather vaguely defined and rather difficult to solve just by using a little bit of common sense in the way that scientists are used to do it. On the other hand, I was of course frustrated by the fact that there were such messy problems. I really, really would've loved to get back to research. I did keep a little bit of contact with physics. There was a group that met occasionally in the evening at Columbia University with people from various areas, but that had something to do with Columbia. That maybe a group of five or ten people meet to have a little seminar on some physics related subject and usually solid state or chemical physics. Elliot Montro was one of the leaders of that.

Kojevnikov:

Were they also scientists working for the military or were they at that time teaching at Columbia?

Herring:

I think mostly working for the military. I don't know for sure.

Kojevnikov:

Did you personal style of research change because of this experience by the terms of the amount of resources you use or the mathematical devices you use?

Herring:

No. My personal style of research in physics that I think was unaffected. But I continued to have a great temptation to take non-physics fields that perhaps were related to physics in some way, but not as basic science. Like the problem of how scientists communicate with the scientific literature and things like that and using operation research methods on those. I spent a lot of time doing things like that in later years.

Kojevnikov:

As far as your approach to solid state problems, did it remain the same?

Herring:

It remained the same, but the amount of solid state physics that I forgot, being without almost without any contact. This little seminar at Columbia was not really much. Being almost without contact with solid state physics for several years really put a lot of things that were very close to my main interest just out of my mind completely. I can tell you the story of that short. After the war was over, I went to my first job at Bell Laboratories.

I was put in an office where I shared an office with an experimentalists, George Moore, who was working on thermionic emission from single crystal tungsten. I'd had some contact with thermionic emission problems at MIT where any number of experimentalists in the field and so on. So I was interested in this man's problems and in particular since he was working with single crystals, interested in how the thermionic work function would vary from one crystal face to another which was something you could measure.

I had vague recollection that John Bardeen had done some work on that that I knew about when I was at MIT and he was at Harvard. John had just arrived at Bell Laboratories and I went around to see him. I asked him what it was that he did and if he could explain it to me so that I could make better contact with Moore's experiments. Bardeen said, "Well, I didn't do much of anything. It was really quite trivial. But Smoluchowsky really picked this up and wrote a nice paper for the physical review." I thought, gee, Smoluchowsky I know him. He's a very good physicist. If he wrote a paper on it I really want to know about it.

So I looked up Smoluchowsky's paper in the Physical Review and I started reading it. I said, "Ah, this is good stuff. This is what I want to know." And I read it. Yes, yes very nice, very nice. I'm very glad to learn this. I read it through to the end of the paper. The last paragraph of the paper said, "The author wishes to thank Dr. Conyers Herring for reading and discussing this paper." Then I remembered not only had I discussed this paper previously with Smoluchowsky, I had refereed it for the Physical Review. I had completely forgotten until I got to that point.

Kojevnikov:

So the paper was published what year approximately? Was it before the war?

Herring:

I don't remember exactly. I would guess probably 1941 or '42 during the war.

Kojevnikov:

That's also the question that I wanted to ask before. In the time after you left MIT and all the way to the war and during the war, what were your contacts among the physicists? Did you keep contacts with Bardeen for example or with Benyier [?] or with any of your classmates in Princeton who were solid state physicists? Did you know whom you met and eventually who you didn't meet?

Herring:

Well, most of them were doing war work in some area that I was not and of course we were all bound by secrecy so we didn't have much contact. I did have a little contact with Bardeen because he was working at the Naval Ordinance Laboratory in Washington. There was an underwater problem that he was involved with where the Germans had a new type of, I guess you'd call it a mine that was proving very damaging to Allied shipping in the English Channels where the water was shallow.

So they could put an explosive on the bottom and ships going over would not be terribly far above that mine. The ship passing over would make a flow of water around the ship, which by the Bernoulli Effect would affect the water pressure and the change in pressure would actuate a detonator on the mine and blow the ship up. A very fiendish thing and very difficult to counter and so I had a little dealing with Bardeen on that. But for the most part I didn't have much contact with any of my former colleagues.

Kojevnikov:

Were you personally close to Bardeen or not ??? ??? ?

Herring:

Well, not as close as I became later. I valued him as a friend and had great respect for his ability. We got along well together, invited each other to dinner now and then in Cambridge he at the faculty club or Bert and I at our little boarding club across the street from MIT. Took one— Hill and I took the trip along with Bardeen down from Boston to New York for a Physical Society meeting traveling on the boat rather than on the train. We had a nice evening on the boat playing cards and things like that. But I didn't really become a close friend of Bardine until we'd seen each other for more of a length of time at Bell Labs.

Kojevnikov:

And will it be a physicist whose work you would try to follow in terms of publications?

Herring:

Of course, yes. You soon get the sense of creative evil is watching.

Kojevnikov:

So who were you watching at that time?

Herring:

Well, Shockley, Seitz, Smoluchowsky. Van Vleck of course was fairly senior man by that time, but I somewhere along the line in those years I read his book on Electromagnetic Susceptibilities, which is a classic even today, a marvelous book. I'm sure there are many others. Nottingham was a superb experimentalist and he did such good experimental work, trained all his students to do very good experimental work. I got the feeling that in thermionic and photoelectric emission and that sort of thing you could believe what you saw from Nottingham's students. You couldn't believe what you saw from anybody else.

Kojevnikov:

Did you feel that somebody was doing a similar kind of a work or someone whom you consider a rival?

Herring:

Well of course Slater's group on solid state molecular theory had people doing band theory and so on. Actually I got out of band theory after this long work with beryllium. Although I had plans for doing some before the war, after the war I didn't take those up very seriously. I just got into other fields.

Kojevnikov:

Do you have any particular reservations about the Band Theory, or what were your reasons for you to—

Herring:

Well I had reservations at first, and I'm not sure just what time I got rid of them. It seemed to me that details of Band Theory exactly how these energy versus K curves went and exactly when two bands were degenerated at a point and things like that were things I concerned myself with in my thesis. I felt that I really wondered does this really have any physical meaning because these are the characteristics of a solution of the quantum mechanical problem in metals of an approximation to the solution only. The approximation being the approximation that neglects the electron-electron interaction except in the main field sense. The electron-electron interactions contributes the order of the correlation energy as we call it. The part of it that is not taken into account in one electron theories is of the order of an electron volt per electron. What sense does it make to understand the band structure to small fractions of an electron volt if you're neglecting things of the order of one electron volt.

Kojevnikov:

Were there any alternatives available at that time?

Herring:

And there were no alternatives available. I didn't feel that I did my thesis completely. My thesis was completely worthless because I realized the whole mathematics could apply to the phonon spectrum as well as the electron spectrum. For the phonon spectrum the assumptions were okay, but I did worry about the electrons.

Kojevnikov:

Did you mention this as far as in print somewhere? Or did you just keep on fixing it?

Herring:

I didn't mention the mix explicitly I guess. But somewhere maybe in the late '40s or early '50s or something. Before Landau came out with his theory of Fermi liquids, I began to realize that at least for electrons near the Fermi energy the quasi particle idea made some sense. The quasi particle spectrum would be similar to the particle spectrum on calculated and self consistent Veo theory.

Kojevnikov:

How did you get acquainted with the quasi particle idea?

Herring:

Just intuition I guess. Just realizing that when an electron is just a little bit outside of the Fermi energy, it's lifetime with a spectra scattering by other electrons...

Kojevnikov:

I was wondering if there were any discussion at that time among the theorists about the applicability of the band theory? Either at the seminars or in the literature or between solid state physicists.

Herring:

The literature was rather confused. I think a lot of people did not appreciate the importance of correlation energy. Wigner and his students of course did, and I fully agreed with them. But I remember one appalling paper by Brillouin who had a big reputation and much of it well deserved in fact, who I think using what's now called the Brillouin-Wigner Perturbation Theory which is a modification of the Schrödinger Perturbation Theory. Published a paper on the perturbational treatment of the electron-electron interaction in metals or in a free electron gas.

Although if you tried to greet that interaction straightforwardly by perturbation theory using the usual Schrödinger type of perturbation expansion, you find that the expansion diverges because of the fact that the coulomb force is so long ranged. In other words if you're doing something in a K space, a momentum space things blow up as K goes to zero. In the Brillouin-Wigner method I think he found that he could get around that, and he could sum the perturbation series to all orders. He found that when he did that, the way he had grouped his terms that some to every order was zero therefore there was no correlation in it. I was simply appalled by the lack of physical intuition that showed. There must be correlation energy and how did he get such a wrong answer.

In looking at what he did, I convinced myself that the mistake was taking a conditionally convergent series and summing all the various terms in the wrong order or taking limits into the wrong order. One limit is the limit of volume going to infinity and one is the limit of electric charge going to zero and so on. So you took your limits in the wrong order and you got zero answer for something that should be finite. So that was one example. Another example lasted much longer than that. Slater himself, although he was quite aware that there was correlation energy, he was not willing to accept the estimates that Wigner and others made of the magnitude of the correlation energy. He was more inclined to emphasize Hartrell-Fook calculations than people who were really afraid of correlation energy. Lud [?] Dean, who used to conduct meetings or summer schools I guess you'd call it down in Florida on molecular and solid state theory. Again, published some papers on binding energies and things in metals in which he-

Kojevnikov:

What year?

Herring:

I don't know; probably the late 1950s or the 1960 something like that. ...In which he got agreement with experiment on binding energies using much smaller correlation corrections and what one got out of the Wigner theory.

Kojevnikov:

In speaking of ??? papers and tje field in general, we come back to the late '30s and '40s. Other than Physical Review, what journals, for example European journals did you try to follow.

Herring:

Of course in the '20s and early '30s, Zeitschrift füz Physik was the great journal. As some of the many best German physicists emigrated, of course that importance decreased, but I think up until the start of the war the Zeitschrift für Physik was still very important.

Kojevnikov:

Were you still getting the journal itself during the war?

Herring:

I don't really know because I never visited a physics lab during the war. But the other German journals, the Physikalische Zeitschrift and the Annalen der Physik were quite important, again in the '30s. And the Physikalische Zeitschrift of Soviet Union had a lot of important papers in it because there were quite a number of Russian scientists who names were very familiar in the west and who were very highly respected.

Kojevnikov:

Did you notice any or did you remember that you noticed any particular work done in Russia at that time?

Herring:

Yes. A lot of the Landau-Lifshitz papers for example like the one on feromagnetic domains and a variety of other subjects were published in the Physikalische Zeitschrift of Soviet Union.

Kojevnikov:

I think Phillip Anderson told me that you also read in Russian.

Herring:

Oh yes. A group of us at Bell Laboratories decided immediately after the war that Russia was not going to publish any Physikalische Zeitschrift of Soviet Union after the war and although they did publish a journal of physics in English, we were not sure how long that would last. We were afraid that it would be necessary to read Russian to get these papers, and so we formed a class. We didn't have a teacher, but we bought the thinnest Russian Grammar we could find and started to study it.

Kojevnikov:

So that's how you learn some Russian?

Herring:

Yes. So I finally got enough of a toe hold in the Russian language to read or to decipher physics papers because the physics vocabulary is rather limited and uses some Latinized words. Gradually as I read more my barrier against reading went down so I read more and more. Soon I was reading the physics papers very easily. Then I started reading things that were not physics and I got so I could read nonphysics at least after a fashion. Then in 1960 I got an opportunity to visit Russia. I said to myself, "Gee you read Russian!" or I guess I thought in Russian, I said, "[Russian phrase]." I found a Russian stouchga [?] in our town, Summit, New Jersey, who gave Russian lessons. I spent evenings with her over a summer and was actually able to converse in Russian then and I gave a lecture in Russian at the Institute of Semi- Conductors in Lenningrad.

Kojevnikov:

You said that you wanted to read Russian. Who else was there?

Herring:

A number of the Bell Laboratories people, not so many of the people whose names you would particularly know, solid state physicists. Allen Holden, who was something of a physical chemist. Very bright. I don't think he had a Ph.D., but he was very intellectual man in all ways was one of the outstanding ones. Willie VanRosebrook, another man who did not have a Ph.D., but was still on the research staff and did some spectacularly good work later. I'm trying to think who else. We had a group of, I don't know, maybe eight or ten people.

Kojevnikov:

What Russian journals would be available at the Bell Labs library?

Herring:

They had the JETP right there. [Russian phrase], of course.

Kojevnikov:

And Doklady?

Herring:

Doklady, of course, yes.

Kojevnikov:

As you were reading at that time, we're talking about late '40s. As you were reading Russian work, did anything strike as a similiarity or some difference in style.

Herring:

A little bit yes. I could see differences in style. Not so much between America and Russia, it was between different groups in Russia. I found that the Landau group did just the sort of thing that I would've expected out of them. Sometimes a little bit long on formalism, but very good physics. The Sverdloskk group, where the main man was Vonsovsky, were much clumsier I thought. They seemed to think that if they formulated everything in terms of second quantization they were really on the frontier and so on. A lot of this stuff struck me as rather trivial. In fact I get I think a completely wrong impression of Vonsovsky as a man just from looking at his papers and finding that most of them weren't very deep at all. But there was so much work going on there that I did pay some attention to it. Then in Lenningrad, of course the elder Gurevich and later the younger Gurevich were very able physicists. Totfe, of course, didn't write papers much at that time, but he was head of the Institute of Semi-Conductors and very knowledgeable over a tremendous area of physics.

Kojevnikov:

Do you still follow any of the work of Frenkel, although he died early? He died in '52 I think.

Herring:

Frenkel? Yes. I was quite familiar with his earlier work, but I didn't follow much of it contemporarily then at that time. One thing I do remember that was relatively recent where I think he missed the boat, he took advantage of a development in the theory of liquids. Namely the idea of treating vacancies as sort of an essential component of liquids and formulating the physical properties of liquids in terms of the properties of vacancies. One of the properties of course that's important for liquids is the viscosity and I think he got that after vacancies. So then he said, "Well, why shouldn't they supply the crystals too? Crystals have vacancies therefore crystals should have viscosity." I shook my head and went no. It is impossible for a crystal to undergo a sheering strain unless you add dislocations in it. Vacancies won't do it. He didn't seem to realize that. But some of his other work of course was quite important and certainly important historically.

Kojevnikov:

Should we maybe now go back to the end of the war and how you made the transition from the war time work to Bell Labs?

Herring:

Yes. Well as I said, I went first to the University of Texas.

Kojevnikov:

Did you consider returning to the University of Missouri?

Herring:

I considered that. I had been left as an instructor and they offered me full professorship to come back. That's the way all academic values had escalated, and at several times the salary that I had had as an instructor.

Kojevnikov:

What was the salary at your military work compared with academics salary at that time? You did the work for the military.

Herring:

Yes. The academic salary was quite good. But Texas offered me a higher salary and somehow I thought it would be a little more glamorous living in Texas, so I accepted their appointment. But I accepted it to start in the second semester, which would be January 1946. So for the last few months of 1945, I took about a three month temporary job at Bell Laboratories.

Kojevnikov:

Who invited you?

Herring:

I think it must have been Dean Wildridge. I interviewed by the mathematics group there under Hendrick Boda, but I didn't want to transfer to just a mathematics group. I wanted to do solid state physics. Bell Laboratories was just reorganizing itself to more of an orientation towards solid state physics. They had a solid state physics department jointly headed by Shockley and Stanley Morgan. The physical electronics group, which studied largely surface properties of solids under Dean Wildridge. I had known Dean Wildridge when we were both graduate students at Cal Tech so I thought that would be a very nice place to go.

Kojevnikov:

Did you go there to work on a particular problem?

Herring:

Not a particular problem, no. Just to join that group. And of course at this time this was only a three month job anyway. But I had a great admiration for Wildridge. I came to Bell Laboratories and I found that they were hiring Bardeen and of course Shockley I knew was already there and a number of other people that were very good at solid state physics. So I thought this is really going to be a good place to work. So when Bell Laboratories, after I had gone to Texas, Bell Laboratories made me an offer of considerably more money than I was getting in Texas to come to Bell Laboratories. And that combined with the fact that the research atmosphere was much stronger too, no question I wouldn't rather leave Texas and go to Bell.

Kojevnikov:

Was there a need for physicists of solid state in Texas?

Herring:

What?

Kojevnikov:

Who was among the physicists in Texas? Who would your daughter—

Herring:

Let's see. I was actually not in the physics department at Texas. I was in the applied mathematics department. They have two mathematics departments, I was in applied mathematics. I was the only physicist in that department, but they did want it to expand in theoretical physics. They asked me did I have any names to suggest to them for people they can hire, and I was appalled at the extreme rejection they showed for the names that I suggested because both of them were Jewish and one was a woman.

Kojevnikov:

Okay. Who were they?

Herring:

Jenny Rosenthal was one and Eugene Feenberg was another, both very good physicists.

Kojevnikov:

Do you think that Jewishness and being a woman was a factor?

Herring:

I was told that a Jewish person would never be approved by the administration.

Kojevnikov:

Were there any Jewish on the faculty?

Herring:

Although that was not told to me by anybody on the administration that was told to me by the leading man in my department.

Kojevnikov:

And there were no Jewish faculty members in the department?

Herring:

Not that I know of, no.

Kojevnikov:

Any women?

Herring:

I should explain though one thing we sort of skipped over that you may want to touch on. And that is in between completing my year at Princeton and starting my year at Texas, I spent the summer at the Ann Arbor Summer School on Theoretical Physics, which they had every year. Which was a very interesting experience and I got the chance to hear some of the leading authorities on various subjects. They'd give their courses of lectures and so on. Like Fritz London and lets see, Fermi was there for a while. I don't if he gave a course of lectures. London gave lectures on his theory, Phenomenological Theory of Super Conductivity, which impressed me very much.

Kojevnikov:

How about hiring Jewish physicists at Bell Labs?

Herring:

Pretty good. Shockley was particularly strong opponent of anti-Semitism. His ideas and his recommendations for people and so on were taken very seriously by the administration. I remember one case in particular after I had been working on semi-conductors for a while. That was a few years later. I had been very impressed by a paper by Hobalist [?] and Mybome [?] on magneto resistance for germanium, of n-type germanium. I heard from Mybome, I guess it was that he would be very interested in a job at Bell Laboratories. And at just that time the word was coming down from top management that we're in a tight spot financially. We just aren't going to have any openings. So I transmitted that bad news back to Mybome, but I happened to mention this to Shockley. And Shockley said, "Well, maybe we can do something for him and got him a job at Bell Labs." He was I guess a Dutch Jew probably who had been, I don't know whether he was in Israel at the time he wrote this paper or not, but something like that.

Kojevnikov:

How did it feel returning to solid state in physics at that time? Did you feel that you missed several years? Did you feel that you had to catch up with some development or had the field changed?

Herring:

Well, there was a lot that I needed to catch up with.

Kojevnikov:

Like what?

Herring:

Well one thing for example that a lot of people got into with both feet, I didn't really get into it, but I remember feeling that this is something I really ought to know more about is the cooperative statistical mechanics. When I was at the Ann Arbor Symposium, I shared a dormitory room with my friend from Princeton Gregory Wannier. And Wannier had been, I guess he'd been overseas just before the war, but he'd come back here early enough. But he had been collaborating with Kramers on the statistical mechanics of the two-dimensional Ising ferro- magnet. And their collaboration, they'd made some important steps together, but before the work came to a really good conclusion they got separated and he lost contact with Kramers completely because the war had started. So he was trying to finish up the paper on his own and publish it with Kramers name on it, but the later part of work would be done all by Wannier. And he made a marvelous discovery of the symmetry relation that exists between the high temperature partition function and the low temperature partition function for that particular problem. Because you have that symmetry, you can find that the partition function on the temperature scale near here are related. Now if you bring these two together, there's some point in the middle where the high temperature result gives the same answer as the low temperature result. And those can be calculated by series, but in conversion way and that then must be the point at which singularity occurs at which the curry point occurs if there's only one singularity. So you have an exact determination at the curry point which is quite an achievement. He published that and that was dually appreciated by everybody. But then I remember in the middle of the war, I was living in New York and Gregory Wannier came around to see me just briefly for an evening. And he pointed out this paper that Onsager had just published that solved the complete problem over the whole range exactly—

Kojevnikov:

Was he himself at Texas at that time?

Herring:

Who had?

Kojevnikov:

I thought Wannier, he had done work in Texas somewhere.

Herring:

I don't think so. I don't think he was ever at Texas.

Kojevnikov:

Because I thought he had some job problems, especially during the war. He worked for some Texas oil firm.

Herring:

He might have been. I know he was with some like an oil firm or something, but I thought it was New Jersey rather than Texas. Anyway, he came to me and he was in a terrible state. He showed this paper of Onsager's, which used very sophisticated mathematics, a whole are of mathematics. He said, here I've got to learn a whole new area of mathematics before I can be active in this field now. He eventually did and of course he came to Bell Laboratories after the war and continued to work in this and got other people interested in it too.

Kojevnikov:

Who brought him in? Who brought him to Bell Labs?

Herring:

Who brought him? I don't know. I suppose the heads of some of the departments. I don't know whether it would've been Boldridge [?] or Shockley or just who.

Kojevnikov:

And also about returning to solid state physics after the war, very often people think about their strategy when they start a field after an interruption for several years. So how did you define? What problems did you want to work on? What things would you pursue at that time?

Herring:

I guess the problems I really worked on were things that I just by accident happen to get an idea in, but of course things are not really accidents. I would get an idea in some field because I had been reading up on it and had been reading up on it because I was interested in And here I think the influence of Fred Seitz was very important. I had realized from the beginning that Fred Seitz knew an incredible amount about everything. He read the literature in great detail and knew all aspects of solid state physics. Which was very wonderful because in 1940 he was able to publish this book on Solid State Theory that was sort of the bible of the field for many years and covered a whole large portion of solid state, not quite all. And I felt that I could never be... I felt that I could not digest the literature rapidly enough to cover so many things in only 24 hours a day. That for me the best thing to do is to pick a field or two or three fields to work in and really know those and try to be really good in them and that's enough. But as the years went by, I began to realize interesting things would come up from all sorts of directions. And I would be more aware of those things the more I was able to keep in touch with a broad amount of literature. So I tried to, although I never succeeded in duplicating the breadth that Seitz could achieve, I tried to be as broad as I could.

Kojevnikov:

In speaking of two or three fields, which fields at that time were you choosing for yourself?

Herring:

Well, I was certainly choosing crystal symmetry and band structure and phonon and maybe electron phonon interactions. Then I had gotten interested in surfaces through my contacts with the thermionic emission people. So I sort of broadened out into more and more and more aspects of surfaces. And actually going to some one of Naughtingham's conferences on surface and electronics. I happen to get in touch with some people who were working on centering, which isn't electronics at all. But it seemed to be doing interesting work and I was enough interested in thermodynamics that I thought well, these people could profit from using more thermodynamics. So I got interested in centering and the geometrical— or what I call the aller-movement aspects of surface phenomenon. I adopted that as one of the fields I wanted to keep in touch with. Of course that got me into not only surfaces, but also into volume transport like diffusion and so on. I found some interesting things there. So it went...every now and then something pretty new would come to my attention that I really wanted to work in.

Kojevnikov:

Was it for the purpose of keeping up with literature that you started the journal club?

Herring:

Yes.

Kojevnikov:

At Bell Labs?

Herring:

Yes.

Kojevnikov:

Didn't have this institution at your previous jobs?

Herring:

Well, in Princeton in I guess 1939-40, somebody had started a journal club there. That's where I got the idea. So after I got back to working at a regular job again at Bell Laboratories, Missouri there wasn't a big enough group to really do a journal club, but at Bell Laboratories there was and so I got one started there.

Kojevnikov:

Who else was in the group with you at Bell Labs at early period? Were you directly under Shockley?

Herring:

No. See Shockley had the solid state department together with Morgan and Wildredge had the physical electronics department. I was in Wildredge's department, which is the reason why I always remain friends with Shockley. People who worked in Shockley's department always— No, not all of them became enemies of his, but many of them did. The ones that had a mind of their own and wanted to go their own way.

Kojevnikov:

But you did work under Shockley during the war time didn't you?

Herring:

Not really, no. In the first place, I was only a part-time member of the group so I really wasn't integrated into his overall plan very much.

Kojevnikov:

Did you interact a lot with him when you were in different departments? I mean scientific interaction?

Herring:

Yes, some. I found him a little difficult to talk with about scientific subjects because he could always think so much faster than I could that even when I was right he could argue me down. But I remember one time, I have to give him credit. He was not a really evil man. He was giving some lectures on transistor physics and so on and lecturing about band structure and that sort of thing and electron phonon collisions. I pointed out to him that the conversation laws of crystal momentum in such collision and so on were based on a very fundamental symmetry problem and didn't depend upon accuracy merely a perturbation theory or something like that. And we had a bit of an argument during his lecture. I finally, I guess after the lecture, but by the next one, convinced him I was right. He gave me full credit for it. In his book he gave my argument and said that this was an argument due to Dr. Herring. I was a little embarrassed at that because I thought this was something that anybody with any knowledge of group theory or any fundamental knowledge would know. I didn't think it was original with me at all, but he put it down that way.

Kojevnikov:

Was there also a difference in style with those who were coming from the group theory approaches, and those like Shockley?

Herring:

Yes, yes. Shockley, because his mind was so quick— It's an awful problem in ordering one's life if one is enormously quicker mentally than everybody else and he had to face that problem. One of his reactions, which was not too good, was something of a contempt for the literature. He didn't try to be terribly broad in reading the literature and tended to discount it or if he read a paper he could usually find a quicker way of doing the same thing. While it's true that he usually was able to do anything better than the way it was done in the literature. It wasn't always true. He did suffer some from this disregarding of the literature. Whereas I always felt, well you learn something from the literature even when it's wrong.

Kojevnikov:

Could you give an example.

Herring:

Well I don't know. One thing I know is he in talking about an alloy with different components and so on, how you can get a compositional current due to a thermal gradient. He felt this was something new that he was contributing. And I said, "Oh this is just the Sorray effect." It's got a name. It's got literature all it's own that I knew about because I followed the literature, but he had never heard of it. I'm not saying that he did anything wrong because of that, although I wouldn't be surprised if occasionally he did.

Kojevnikov:

In the physics electronics group, were you the only theorist?

Herring:

No. There were others from time to time. Most of the theorists though were in the solid state group. Charlie Kittel for example. I think Wannier was in physical electronics. Departmental boundaries didn't really matter much for anything.

Kojevnikov:

How about the division of topics or how did the two departments—

Herring:

They had different experimental programs. The physical electronics group had gas discharges or secondary emission, photo electric emission and so on. The solid state group had a much wider variety of topics. A lot of the things I worked on were in their area or even farther into the field. Nobody ever said you work on this.

Kojevnikov:

How did you feel? What were your obligations towards the department or whether somebody's blaming you? What were you expected to do? What was your job assignment or whatever, expectations of what you would do there?

Herring:

Well, one thing I would talk to people.

Kojevnikov:

You mean experimentalists?

Herring:

Particularly experimentalists and if they had a problem I would try to see what might be the answer to it. And actually many very interesting things came out of that. One of the most interesting was my talking to Ted Gibald [?], who's here at Stanford now. He was in the chemistry department, but he was doing essentially physics. He was measuring the thermal electric power of germanium down to low temperatures where it hadn't really been measured before. And he found that at room temperature the thermal electric power was just about what the theory would predict, a nice level of thermal electric power independent of the temperature. Then as he went down to low temperatures, changed the order of the magnitude at low temperatures and what could be the cause of them.

I resorted to some thermodynamics, and I realized that since the thermal electric power is related to the Celtiea [?] coefficient by thermodynamics, that would mean that that electron is flowing in an isothermal conductor, would carry a tremendous amount of heat along with them. Why? And it couldn't be the electron band motion. That certainly is of order of thermal energies and no more relative to the Fermi level which is the middle of the gap. Of course that can be sizable, but it doesn't change all that rapidly. We know how the Fermi level behaves from the other electrical properties. So if the electronics can't be carrying, what else can carry heat in a semi-conductor? Well phonons. And this led me to think of the whole theory of phonon drag thermal electric power, which became a major field. Actually I found that the elder Gurevich in Russia had proposed a phonon drag theory for metal way back in 1946. But what is in his paper, you could never explain what happened in germanium where it increased so rapidly to those temperatures. Because he treated all the phonons as a sense of the equivalent having a certain push on the electrons related to the heat current that phonons were carrying in a thermal gradient. I pointed out that the amount that the phonons were off of thermal equilibrium due to being pulled by the electrons in a constant negative would be very dependent on the wave number of the phonons. Those little phonons they get pulled very far off the equilibrium and that was the cause of the big effect. And so it went.

Other things were quite different. I remember J.B. Johnson, famous for Johnson noise, showed me some secondary emission data that he had which showed very peculiar behavior, which I was able to explain by means of simple space charge theory. You put in some incident electrons for an external source and if there's a space charge there already from the thermionic electronics, that space charge field will be modified by the field of the incoming electrons. And what you measure as an apparent secondary emission is really a space charge effect and so it went. So all sorts of interesting things popped up at unexpected times. Another one that was very cute was, actually where I had my name on an experimental paper although I didn't do the experiment. Some of the people who were actually working on instillation of telephone devices in the telephone world were finding that a certain gadget that they had in a telephone system was developing short circuits. They took it out and examined it and couldn't find anything wrong with it. They put it back and it worked fine, but short circuits continued to appear here and there. And finally somebody after opening one of these things open, just happened to get the light on it right.

And so a glint of light, which on closer examination turned out to come from a little filament of metal that had grown out from one of the electrodes going all the way across to another electrode a couple of millimeters away and made it short circuit. In fact the whole electrode had developed sort of a set of whiskers. This was reported in Bell Laboratories news just as a piece of general news for all Bell Laboratories people. Some of the people in the chemistry department got to looking at these whiskers under microscope or an electron microscope even and found the diameter was of the order of one or two microns. I thought, gee something this small might well be a single crystal and a perfect crystal. The dislocations and things might easily have moved out of it because they don't have very far to move to get to the surface. Since the theory of plasticity, which I was somewhat familiar because I had taken to reading Baud's portion of the literature, the theory of plasticity said that a really perfect crystal should withstand an enormous strain without deforming because it would have no dislocations. So why don't we try to put an enormous strain on these? So John Gault got a micro manipulator and a microscope and bent some of these whiskers around. They took a percent or so of strain and jumped back into their original form when you took the pressure of them, no sign of any deformation. That started a flurry of studies in the literature of the elastic and plastic properties of whiskers. So it went. People talked to you and suddenly an idea pops up.

Kojevnikov:

I'm wondering how did it feel moving to an industrial lab from the university? Did you feel that the moves were somewhat different there.

Herring:

It's like what a university is supposed to be aside from teaching. Bell Laboratories achieved the ideal of the research university better than any university could.

Kojevnikov:

Did you show up at work everyday?

Herring:

No. Most days I went in, but a fair number of days I would stay home all day and work at home.

Kojevnikov:

Just kind of finishing up, when you were finally leaving Bell Labs, I'm jumping a little bit, for the university what were the reasons for that? Or why would you choose an ideal research university for—

Herring:

The long time policy at Bell Laboratories was everybody retires at 65. And while there was some talk of possibly changing that and so, and it did get changed. I wasn't sure at the time, I don't know I was in my early 60s, 63 or something like that 64, I wasn't sure that that would really be changed enough for me to continue on there indefinitely. And I thought well if I go to a university even if they have a compulsively retirement, I can be a Professor Emeritus. And a Professor Emeritus usually has an office and circulates with the other professors and he can do what he pleases so maybe I'd do better to go to a university. So I considered possible university positions and got a particularly good offer from Stanford for three years as professor and then retiring. So I took that.

Kojevnikov:

Do you have in mind to mention any specific anything? A third topic today or should we—

Herring:

No, nothing particular and more I guess.

Kojevnikov:

Then maybe we should stop for a while.

Herring:

Yeah. [pause in tape]

Herring:

At the start of my work, some of the U.S. scientists had visited England, looked at a lot of documents and taken microfilms of documents having to do with subsurface warfare from the British files and brought them home intending to blow up the microfilm and have it photocopied and distributed to all our laboratories. Well when they tried to blow up the microfilm, they found the quality of the microfilm was such that it would not be legible if it were blown up. So they decided the next best thing would be to have somebody who knew physics look at the microfilm on a microfilm viewing machine and dictate what he read to a stenotypists who type it and then get it transcribed. So here I am, I know nothing about subsurface warfare. I'm given this tape on subsurface warfare, or film rather to read to a stenotypist who knows even less to be transcribed by still a different person. I finally saw the results that came back and they were incredible. Errors, errors, errors every place you could imagine.