John Bardeen - Session IV

Hoddeson:

This is Lillian Hoddeson and this is Session 4 of an oral history interview with Professor John Bardeen in his office at the University of Illinois. It's December 22, 1977. We left off last time soon after your arrival at Bell Labs in late 1945. To help refresh your memory of what Bell was like for you in those days, I brought along an organization charge, several in fact.

Bardeen:

Was it 1945 or 1946?

Hoddeson:

This chart is the July 1946 replacement of the January 1946 chart and on it we see the three new groups that had recently been formed under Wooldridge, Fisk and Shockley. For Bell, as I understand it, the formation of those groups was a departure.

Bardeen:

Yes, (looking at the Morgan-Shockley group), this is the new group which was formed after the war to work on solid state physics. I think I said earlier that Kelly and Fisk were probably the main instigators for this group, and Shockley played a big role in setting it up. It was Fisk and Shockley who got me interested in coming to the Labs, and according to this chart, was at that time in the group on semiconductors when I first arrived in late 1945 -- October or November 1945. I didn't know what field I'd be working in and as I think I said earlier, they put me in an office with Walter Brattain and Gerald Pearson and I started work on semiconductor problems. As you can see on this chart, I'm in the semiconductor group.

Hoddeson:

Were you aware of the master organization as reflected on this chart at the time? When they put you in that office, did that appear to you as though it was more or less accidental?

Bardeen:

This is July 1946 issue replacing issue January 1946, so at that time, this group was already formed. But I doubt if was formed in January, and it certainly wasn't when I first arrived. Because I looked over the activities of different people in magnetism and other areas and then decided to work on semiconductors.

Hoddeson:

Kelly had been talking, since the mid-thirties, about setting up a multi-disciplinary solid-state group.

Bardeen:

Yes. He wanted to get a group of both chemists and physicists involved.

Hoddeson:

Yes, he had been talking about that for years but he wasn't in a position to get it organized until this time.

Bardeen:

During the war of course, he couldn't, but right after the war, he got it started, and many of the people were working on various wartime projects during the war including Walter Brattain and Gerald Pearson.

Hoddeson:

Yes. It was also through their work that Shockley got interested in some of his work on semiconductors before the war.

Bardeen:

Yes, he'd worked with Brattain on semiconductor problems before the war, but neither one worked on semiconductors during the war. But this is the group set-up here, which is essentially the same group we had when we discovered the point-contact transistor. About the only additional people involved were those in Goucher's group, who were then transferred to work on semiconductor problems; the group under Shockley was augmented by two groups under Goucher.

Hoddeson:

Did you work closely with Goucher's group during the transistor work?

Bardeen:

I didn't, but Bill Shockley did. There was the famous Haynes-Shockley experiment. Some of these were Technical Assistants, like Ryder and Fois. (J.R. Haynes and W. Shockley, Phys. Rev. 2. 691 (1949)].

Hoddeson:

And Griffith?

Bardeen:

I guess he much have been a technical assistant. Gibney was a chemist; Moore was an electrical engineer. I see Townes is listed in this group but I'm sure he didn't stay there very long because he was in a different area, not in solid-state physics.

Hoddeson:

Well, he worked under Morgan.

Bardeen:

Yes, he's listed under Morgan. But, I think he was working with a different group at the time, not in this but the solid-state physics group. I'm not sure which one.

Hoddeson:

I wish I had the previous chart which reflects the war organization. On the very next chart, Townes is still under Morgan, and then in 1948, I don't see Townes at all.

Bardeen:

1948 was, I think, when I worked with him. He was working on microwave problems, the spectra of gasses.

Hoddeson:

Perhaps we should take a minutes to discuss the work that you did with Townes and we will be able to focus entirely on the transistor. I'm interested in how you got involved with this work and how it related to other problems at Bell.

Bardeen:

Townes came around with the problem which he had been working on, the spectra of molecules, and he could make very accurate measurements, and the problem was to calculate the effect of quandrupole moments and hyper-fine splittings. I worked out the theory of the effect, and some of then complicated looking formulas here that just involve quantum numbers and we could calculate the position of the lines very accurately to many significant figures. We had a real feeling that quantum mechanics was correct since we could calculate the position of all these lines to many significant figures.

Hoddeson:

What did this have to do with Bell Laboratories' interests?

Bardeen:

I think he got interested in this because he worked on microwaves during the war and then this was an application of measuring the microwave absorption spectra of gasses. We could measure the lines very accurately and he noted that you could determine the nuclear quadrupole but he needed a theory to work out and analyze the experiments. And, I worked out the theory.

Hoddeson:

And the Bell Labs' research group was set up in such a way that he could just walk in and way, "Well, Bardeen, I need some help with this calculation. Do you think you can find the time?" That was perfectly O.K.?

Bardeen:

Oh sure.

Hoddeson:

You didn't have to go to Shockley or anyone?

Bardeen:

No, no third party.

Hoddeson:

I am interested in the way in which the solid-state group was set up. It was set up as an interdisciplinary group to solid-state physics divided up into magnetism, crystals, die electrics, semiconductors. I was wondering whether the organizational attempt t put people working on different aspects of solid-state physics all together in a formal group in practice actually led to interdisciplinary work.

Bardeen:

Well I think it did. We worked very closely together at the seminars where we reviewed earlier work.

Hoddeson:

Are you talking abut the weekly solid-state seminars?

Bardeen:

Yes, all of these people would attend the seminars. Well not all, but those interested would. We had close contacts with Gibney for example. He was a chemist and we interchanged ideas with them, and also with other people in chemistry and metallurgy who weren't in this group. We had close contacts with them too.

Hoddeson:

You're suggesting that the main interaction among the disciplines came in the seminar discussions.

Bardeen:

Or just in informal discussions at any time, discussions at lunch for example.

Hoddeson:

Were all these sub-groups of the solid-state group physically working in the same general location. Was everyone in the Morgan/Shockley group at Murray Hill?

Bardeen:

Townes was not. I don't think he was in this group very long. His office was in a different place with his laboratory.

Hoddeson:

I see. Was Kelly around a lot? W as he sort of in the background and watching these interactions going on?

Bardeen:

We'd see him occasionally, but not too frequently.

Hoddeson:

Would he talk to you about your work?

Bardeen:

No. We talked with him about the work indirectly.

Hoddeson:

The mythology is that he was the organizational mastermind behind this solid-state group and I suspect that certainly true in part. I was wondering how he functioned.

Bardeen:

He was too high up in the organization to worry about people down at the bottom. He was certainly instrumental in setting it up, and I'm sure followed what went on through the reports and things like that but not directly. Only when we had something significant to show him, then he might come around to visit occasionally, but not very frequently.

Hoddeson:

Let's stay with the seminars for a minute. Now Herring recalls several seminars: a solid-state seminar, and then a similar seminar that his group attended, let's see, he isn't on this chart yet but he will be on the next chart; he appears in the Wooldridge group which was later taken over by Addison White. Herring says they went through a similar seminar going through DeBoer's (J.H. DeBoer, Electron Emission and Absorption Phenomena (N.Y. Macmillian, 1935)] book on the electronic processes in solids just as the solid-state group was going through Mott & Gurney's [N.F. Mott and R.W., Gurney, Electronic Process in Ionic Crystals (Oxford, 1950)] text, and others.

Bardeen:

Well, we went through a number of texts and the people involved would be different. I was interested in semiconductors and went through the work done during the way which was just published at that time by Torrey & Whitmer H.C. Torrey and C.A., Whitmer, Crystal Rectifiers (McGraw Hill, 1948)].

Hoddeson:

Did you go through Torry & Whitmer's book in the seminars?

Bardeen:

No, not the whole book in detail but the parts we were interested in. And then we went through Shottky's papers which were published just before the war.

Hoddeson:

Had you, by the way, seen them before the war?

Bardeen:

No.

Hoddeson:

Shockley had of course.

Bardeen:

Some later ones were published in 1941. I'm not sure whether we had access to them or not.

Hoddeson:

Shockley I know saw the earlier papers; I think there was a 1939 paper.

Bardeen:

Yes the earlier papers, he was familiar with those by Mott and Shottky on the rectification. But there were later papers on the metal semiconductors rectifier in which Shottky went into theory in much greater depth, some written with an associate of his, Spenke. That's one of the things we went through; we were trying to get caught up in what was going on in semiconductors. As you say, we went through Mott & Guerney and Pauling's book on the nature of the chemical bond. And I've forgotten what the other books were.

Hoddeson:

Who chose the text, or the papers to be reviewed?

Bardeen:

Just general agreement, if we had enough people together interested in going through them or participating in seminars. It was very informal; we jut got together and did it. As I recall, at first they were out-of-hours. We got together at five o'clock after the official quitting time. Then later, we did it in-hours.

Hoddeson:

Were they once a week meetings?

Bardeen:

At least once a week, sometimes more frequently. I think in the semiconductor work it was more frequent.

Hoddeson:

I see. And this was in addition to the regular colloquia. Now the Journal Club began at some point also, was that later? I have not been able to find out when the Journal Club began.

Bardeen:

I don't remember now.

Hoddeson:

Herring has some records but they begin in the fifties.

Bardeen:

I'm sure it began before that, just when I don't know.

Hoddeson:

And then Herring remembers some formal lecture series in quantum mechanics and statistical mechanics. Did you attend those? They may have been on a more elementary level.

Bardeen:

No, I didn't go to any of those.

Hoddeson:

The main series then was this solid-state series, the informal going through papers and texts that you were involved in?

Bardeen:

Yes, there was that and then the Journal Club in which we'd review the current literature. We even got a group together to study Russian, to learn to read it anyway.

Hoddeson:

Before we leave the seminar, I'd like to ask another question. The people who were involved had very different backgrounds: Brattain had essentially no quantum mechanics, Gibney is a chemist, and so on. The general level of the seminar must have been very variable because the people who were involved were so different in their training.

Bardeen:

Well it would depend. For example, going through Pauling's book was right up Gibney's alley and Mott and Gurney also did a good bit of chemistry. We did more chemistry than quantum mechanics; we used essential ideas from quantum mechanics but not detailed theory. I think that was true of Schottky's papers too, once you accept the basic ideas of electrons and holes, it's essentially classical ideas. The things we talked bout in the seminar didn't require any deep knowledge of quantum theory.

Hoddeson:

Did you also discuss the reports that were coming out of the Purdue group? For example, on germanium?

Bardeen:

Yes.

Hoddeson:

They were available?

Bardeen:

They were available. I don't know how much of it we went through in the seminar. I know they were available. Lark-Horovitz and Johnson looked at the transport properties of semiconductors in considerable detail and went through those papers. I think they were published just after the way or in the years just following the war. We also had the wartime reports available. So we were pretty well up on semiconductors. As I said, it was a new field for me, so I was learning from any source I could use.

Hoddeson:

You never read this stuff before the war then? Mott's theory of rectification for example, had you read that before the war?

Bardeen:

I read that before the war, but it was just general reading and nothing else. I had no interest in working on that on my own. But Schottky's work was new, that's true, later papers were new.

Hoddeson:

Were there any other important pieces of work going on besides the work of people like Schotty, Mott and Davidoff who were writing theoretical papers and experimental work such as by the Purdue group? I'm trying to learn what were the main pockets of work that people were trying to learn at that time to gain a general background.

Bardeen:

Well, I would say Torrey and Whitmer, which told what they did at the Radiation Lab during the war in semiconductors.

Hoddeson:

Did they do much?

Bardeen:

They did a good bit involving silicon for detectors and we went through theory of rectification. I think Esther Conwell and Weisskopf did that. Some theory of scattering. There was good bit of work done during the war which we took advantage of. And Schottky's work and then during the war, the concepts of doping were rather vague. It was thought that you could dope with three to five compounds and get p-type, but it hadn't been proven, say that a group five element really did go in such extra electron could be detached from the donor and make the semiconduction in type. It's one of the things Gerald Pearson was working on, to try to pin down these ideas more definitely, and he made a good many measurements on transport properties. He and I wrote a joint paper which was published in 1949, I think.

Hoddeson:

Oh yes, that beautiful silicon paper. I have that with me.

Bardeen:

It was a very popular paper after the transistor came out for people who wanted to get into the field. I've heard that Bell Labs issued something like 5,000 reprints, so that paper probably got distributed more widely than any other that I've written.

Hoddeson:

It has a very beautiful balance of theory and experimentation.

Bardeen:

Most of the ideas came from earlier work. This was really putting them all together and showing quantitatively how things worked out.

Hoddeson:

Let's see, during the war you mentioned the theoretical work, work of the Rad Lab, work, of the Purdue group and Bell. Now there was some work done at Penn, wasn't there?

Bardeen:

There was some work done at Penn. We had all those reports available to us. The work at Bell was mostly in metallurgy and materials. One was working at Holmdel. He was one of the first to suggest using silicon as a detector for microwaves and he did experiments on silicon. Brattain was familiar with that work and some of the data before the war.

Hoddeson:

Did you interact much with Ohl?

Bardeen:

He went down and talked wit him at Holmdel on quite a few occasions.

Hoddeson:

He seem to have been fairly close to something like a transistor himself. Do you feel that is true?

Bardeen:

He observed some of these optical effects, photo-effects which suggested that there was in inversion layer on silicon.

Hoddeson:

Apparently, he built a radio without using vacuum tubes, using negative resistance elements, and that was demonstrated just after the war to Shockley. Shockley mentioned it in one of his articles. Does that ring a bell?

Bardeen:

I don't remember that it is quite possible. Using thermistors?

Hoddeson:

Yes, I think so.

Bardeen:

Using a diode for detector.

Hoddeson:

Yes, but it was apparently very unstable. I left a strong impression on Shockley. I don't have any documents to back that up.

Bardeen:

He was a sort of intuitive experimenter. He didn't have any great knowledge of theory but he did these experiments.

Hoddeson:

In what language did you two speak to each other? Ohl didn't know quantum mechanics. I gather he used a sort of chemical language.

Bardeen:

We just talked about the experiments. Just talked in terms of what experiments he was doing. The ideas of n- and p-type or excess and defect, started being used during the war. We talked about it in a natural language. He was able to prepare a surface such that it had an inversion layer on the surface which Brattain remembered. That's one reason we started with silicon rather than germanium. The first electrolytic transistor was with silicon. That was the reason for picking that particular combination. We had some evidence from Ohl's work that there was this inversion in silicon and we made use of the inversion layer.

Hoddeson:

Did you understand that that was an inversion layer when you first began working with silicon?

Bardeen:

Well, I suppose we're getting a little bit ahead of the game, but when Brattain and Gibney found a change in contact potential with light and that you could move the surface barrier up and down by applying an electric field through an electrolyte, they thought that with their experiments, as I remember, were with both silicon and germanium. The idea of using a point contact geometry was mine, to try to make an amplifier. And the idea for doing that was to try to get away from thin films which we had been using before but to use bulk materials. This experiment they were doing of course used bulk materials, and bulk material has much better electrical properties, much higher mobility, so that part of this was just a simple way of testing the idea.

Hoddeson:

I have a question here. In looking at the series of experiments that led to the final contact transistor, I see you started out with the field fact and then the effect of moisture led to replacing the electrolyte with deposited thing oxide film and then...

Bardeen:

It was just trying to observe the field effect.

Hoddeson:

It was only at the end and this is the question I have -- it seems almost to have been by accident that you went to a point contact effect....

Bardeen:

We went to the point contact to observe the field effect and that was why we insulated the point from the electrolyte. We accepted the field across the electrotype.

Hoddeson:

At what point did you realize that holes were coming through?

Bardeen:

Well, that was later. Experiments worked with the electrolyte but we wanted to get rid of the electrolyte and tried to do that by forming an oxide and then evaporated gold on the oxide, and then put a point contact across to observe the field effect. We found an effect but it was in the opposite direction than you'd expect from the field effect which showed something else was going on. And, that's when we thought this must be injecting holes.

Hoddeson:

That late along in the series?

Bardeen:

We just found it experimentally first and it was just a small effect. The gold contact wasn't insulated from the germanium as the thin layer of oxide on the surface wasn't enough to insulate it. It's like the geometry of the present MOS transistors. This wasn't discovered until I guess the early sixties -- how do we make a good insulated oxide so you can make MOS transistors.

Hoddeson:

The turn in this series of experiments that led to a transistor seems to me to have occurred when you noticed that water was interfering with an early experiment which Brattain conducted just after your surface state theory came out.

Bardeen:

After the surface state suggestion, Brattain started work on surface problems and Pearson was involved in studying bulk properties.

Hoddeson:

But in the process of studying the temperature dependence of the contact potential, Brattain noticed a hysteresis effect due to water. And, that's what led to immersing the apparatus in an electrolyte. This is according to Brattain's description. I mean I'm getting this from his description.

Bardeen:

I don't remember details of that.

Hoddeson:

It struck me that while they were just trying to get rid of the hysteresis at first, in fact the water was doing something very important. I don't know if that was recognized at that time.

Bardeen:

I don't remember. There was the discovery that you could do measurements of the change of contact potential with light -- that you could move the surface barrier up and down with an electrolyte which showed that you could get around the surface states if you applied a strong enough field through the electrolyte so that you could get by the surface stage.

Hoddeson:

The question is what did you think was happening with the electrolyte? What was enabling one of get through the surface states?

Bardeen:

The ions were very close to the surface and there was a strong field at the surface. Practically all the voltage drop -- the voltage to the electrolyte -- would occur right across the interface of the electrolyte and the semiconductor. So you get a much stronger field as you are applying the voltage over atomic-like dimensions than over finite distance. And so we thought that the reason why it was so successful was that initially we had a lot stronger field at the interface but no doubt the interface itself, the surface itself, was changed since the electrolyte dissolves the surface. It dissolved whatever oxide was there.

Hoddeson:

I'd like to backtrack a little bit. We left out a few things. In looking at what was going on at Bell in the thirties, before you arrived there, it seemed to me there were three major threads that were coming together at the time of the war. The whole thing was interrupted of course by the war. First was the work that was being done by Ohl along with the metallurgists on silicon. Second was the work of Brattain, Pearson and others on the various "isters"...

Bardeen:

Ohl's work was stimulated by trying to get a detector for microwaves which would work in the microwave region. The work on microwaves started independently at Bell when they were just trying to punch the radio frequency range out into the microwave range.

Hoddeson:

That's also what got Bell involved in radar. Isn't it? The fact that they already had people working on microwave frequencies.

Bardeen:

Yes, the fact that they already had people working in this area was one reason they got into the radar problem.

Hoddeson:

The third line centered about Shockley, Nix and Wooldridge -- the new group that was put together and given an unprecedented amount of freedom to investigate problems from a very fundamental point of view.

Bardeen:

I think that one of Kelly's ideas in hiring Shockley was to get some modern ideas on solids introduced into the Lab.

Hoddeson:

Well, apparently Shockley was introduced by Kelly to the idea of the need for a solid-state amplifier very early in 1936. Kelly spoke to him about that need. And, Bell didn't know very much about how to go about getting such an amplifier.

Bardeen:

I don't think there was any concerted effort to try to do it. People were stimulated to thinking about possible ideas and how it might be done but there was no program set up to do it.

Hoddeson:

They all did this business with trying to put a grid in a semiconductor. Shockley tried it. Brattain and Becker tried. They did a calculation which showed that it was unlikely that they could succeed, but they were certainly thinking along those lines. Shockley also attempted to make an amplifier with Holden employing the piezoelectric effect. It didn't work.

Bardeen:

They tried quite a few things which didn't work before the war. Just trying to do something analogous to a vacuum tube by putting the grid in the space charge layer in the rectifier -- the metal semiconductor rectifier -- but the dimensions were too small.

Hoddeson:

Did you do something along these lines?

Bardeen:

No, it was Shockley and Brattain. As I understand it, Shockley would come up with these ideas and Brattain would try them out even though he wasn't very optimistic about the results [laughter]. And he tried out a number of things but none of them worked.

Hoddeson:

Is there anything more we need to say about the studies that you and the other members of the semiconductor group participated in right after the war. I guess everybody studied Wilson (A.H. Wilson, Proc. Roy. Soc. (London) A 133, 458 (1931); A 134, 277 (1931)]. That was a classic paper. By the way, when did you first meet Wilson. I'm just curious.

Bardeen:

I don't remember. I don't think I ever met him before the war in this country. I think he went into business of some sort after the mid-thirties and was no longer involved in science.

Hoddeson:

Was his 1931 theory of semiconductors very widely read?

Bardeen:

Yes, his work kind of set the stage for understanding in terms of excess and defect electrons and holes so this is one of the basic papers, and Frenkel papers.

Hoddeson:

Yes, on photo-conductive phenomena...

Bardeen:

... explanation of the change in contact potential with light, which really involves all the present ideas which were required for the transistor. If you generate electrons and holes at the surface, and the diffusion efficient is different, so that the electrons tend to drift faster than the holes by diffusion, then the electric field has to be set up to equalize the flow of electrons and holes, so there's a change in the surface potential which can be observed as a change in the contact potential with light. So, this involved the ideas of diffusion and mobility, flow in an electric field, and compensation of electrons and holes and also had non-equilibrium conditions in that you had excess electrons and holes introduced by light. And the equations used to analyze these experiments were true in the early thirties and everything needed to analyze the junction transistor. All the basic equations were there.

Hoddeson:

When did you first read Frenkel's paper?

Bardeen:

I think it was in this period after the war that I read these papers.

Hoddeson:

Were they readily available in this country?

Bardeen:

Yes, probably I had to get a translator from the Russians. They were among the classic papers. Before the war, Brattain had been concerned with copper oxide rectifiers and he'd had some data in his notebook on the oxidation of copper, which he analyzed in terms of the theories developed by Wagner and Schottky concerning the defects, vacancies and interstituals in the oxide? But in this case, it was necessary to understand what was happening during the drill and this is one of the first papers which we wrote after the war and I think it's the only paper ever written with all three of our names on it.

Hoddeson:

Is it?

Bardeen:

I think it's the only one with the three names.

Hoddeson:

This is the Journal of Chemical Physics Vol. 14 paper [J. Chem. Phys. 14, 714 (1946)]. Well let's see how Brattain's notebooks were filled with observations. Was he aware of the theory of Frenkel and others....

Bardeen:

I'm not sure of how much he was aware of the theories. This is a period in which--there is a reference to Mott and Gurney which is one of the things we looked at. So I'm sure we studied this sort of thing in our seminars. The basic sort of theory of vacancies and interstituals, in this case copper ion vacancies.

Hoddeson:

According to note #9 in the article, you did the theoretical interpretation and the experimental work was done by Bardeen and Shockley.

Bardeen:

Well Shockley's name got put on it mainly because he was director of the group and we consulted him occasionally about it, but he really didn't have a whole lot to do with it. It was mainly my talking with Walter Brattain about his early data and then developing the theory of it.

Hoddeson:

Was this done before the decision to study silicon and germanium?

Bardeen:

No, we went very early, even before that, to concentrate on silicon and germanium but it takes time to set up and get experimental data. And this data was available, so I decided to work on this. It involves, as you can see in this case for ions, the same sort of equations which are now used to discuss the flow of electrons and holes in semiconductors. The equation of motion for defects in semiconductors, ionic crystals, are very similar to sorts of theory required for electrons and holes. The theory as regards defects goes back to Wagner and Shottky and we discuss here Wagner's theory of oxidation.

Hoddeson:

Is Wagner one of the leading people whose papers an historian should be studying?

Bardeen:

Yes, he did a lot of the pioneering work. Siemens Company in Germany gave money for a Shottky's prize in physics a few years ago. They honored Shottky on that occasion. I guess he was probably past 85 then. He was honored then at a symposium. I gave a talk in connection with the introduction of the prize and Wagner was present on this occasion. I reviewed the importance of Shottky's work. And Shottky, as I understand, was reluctant to have the prize named in his honor, because he thought that Wagner didn't get enough of the credit for the idea of Shottky defect. He felt that more credit should go to Wagner and he was apparently reluctant to have the prize named in his honor because he felt so indebted to Wagner for many of the ideas. Wagner is certainly one of the key people. I'd say the other key people were Wilson, Mott, Shottky and Frenkel.

Hoddeson:

And what about Davidov?

Bardeen:

I wouldn't say his ideas were so basic. He was trying to understand the pn-junction and wrote down some of the equations but he didn't use the right boundary conditions so he didn't get the transistor effect.

Hoddeson:

Was there work going on in European experimental laboratories as well? Most of the theoretical work was done in Europe before the war.

Bardeen:

In this field most of the theoretical work was done in Europe, although Wagner was at MIT. I'm not sure when he came there. He was a professor at MIT for many years. Before the war, work on semiconductors was done at industrial laboratories and was closely tied with applications.

Hoddeson:

Was any experimental work done in Germany and England on semiconductors. The theory was coming out of those countries.

Bardeen:

There was more done in Europe on semiconductors before the war. But there was not a very close correspondence between theory and experiment, mainly because the materials people studied were copper oxide and selenium which were useful in making rectifiers or for photo effects ad these materials are hard to prepare and control the composition, and so they weren't very well suited for fundamental studies. So these ideas we've been talking about are mainly used to get a qualitative understanding out of the quantitate various phenomena. It wasn't until 1939-1940 that rectification was understood. At one time it was thought to be a tunneling effect.

Hoddeson:

You mentioned to me once that you took a trip with Shockley in 1947....

Bardeen:

Summer of 1947.

Hoddeson:

You visited some of the European labs. I was wondering what you learned, particularly whether they were doing anything like what we were doing here on semiconductors.

Bardeen:

We were interested in solid state physics in general.

Hoddeson:

Did Bell send you?

Bardeen:

Yes, Bell sent us to find out what was going on.

Hoddeson:

Was Kelly involved in sending you?

Bardeen:

I think he had something to do with it. He may have instigated it. It was a very interesting trip. It was not long before that Bell reached some sort of agreement with Phillips which allowed us to visit Phillips Labs. I'm not quite sure just what that agreement was, some patent exchange or something of that sort. We were welcomed in the Phillips Labs and we visited Mott at Bristol.

Hoddeson:

Did you discuss rectification with them?

Bardeen:

Yes. I'd done some work on frequency dependence of the impedance of a metal-semiconductor rectifier and talked abut that with Mott and some of his students.

Hoddeson:

And what else did you learn?

Bardeen:

I think Mott, at that time, was more interested in plastic properties of metal, dislocations, and things of that sort. Dislocations hadn't been seen at that time, but we were shown experiments of Runier in Paris which were suggestive of dislocations. I'd have to look up my notes on the trip to find out the details but the highlights I remember. We visited Mott in Bristol an came back on the train to London with Sasnowski who had been picked up by the British just after the war from a concentration camp brought to England and he was working in England at that time but, then he eventually went back to Poland to work on semiconductors. He was head of an institute there. We visited Cambridge. I think we visited Oxford and then we went to Holland and visited Phillips, Casimir and....

Hoddeson:

Were they doing semiconductor work in Holland at that time?

Bardeen:

Yes they were doing very interesting work, not on germanium and silicon but on things like cadmium sulfide which are difficult to dope because you can introduce defects very easily into them. In general, if you try to dope, the charge is compensated by electrons and holes, but by the charge in the defects. They did very nice work on showing how the defects compensate the charge and the equilibrium equations which determine that. And then also at that time, they were working on ferrites, and just beginning to get some good material for ferromagnetic insulating. They had very good people working on that project. We went to Paris and met Neel and we talked about ideas for which he later won a Nobel Prize on anti-ferromagnetics and how you could make permanent magnets by having very small particles so that you only have one domain for each particle and you have a very high coercive force. It was a very hot day in Paris. I think it set the all-time records for Paris on that year for that day of the month; it was a very hot day. We sat at an outdoor cafe and talked with Neel about these problems. We spoke to him in our high school French since he didn't want to speak English. It was a very interesting conversation because he was full of all these ideas about magnetism. And then we went to Zurich where Busch had been working on semiconductor problems. This was one of the few places which had been working specifically on semiconductors.

Hoddeson:

What was he doing?

Bardeen:

He was studying transport properties and the properties of semiconductors. I've forgotten now what materials he was working with but, as I recall, it was not silicon and germanium but other compound semiconductors. He had a very active experimental program. We didn't go to Germany because conditions were still pretty unsettled there. We went to Germany later -- met Shottky and Spenke.

Hoddeson:

When?

Bardeen:

I think that was 1953.

Hoddeson:

Oh, much later. You mentioned earlier that in Europe, theoretical and experimental work was separated. Was the experimental work in the United States by comparison more closely integrated with theory?

Bardeen:

Typically they had different institutes with a professor in charge of the institute and very little interaction between the different institutes. In Göttingen for example, this was on a later visit with Hilsch and Pohl, who did very fine work before the war on ionic crystals.

Hoddeson:

They actually succeeded in placing the grid into an alkali?

Bardeen:

Yes, for a very low frequency. At that time, Hilsch was interested in thin film superconductors at low temperatures. Pohl had retired but still carrying out some research on ionic crystals. Pohl was a typical professor type. When we talked with him about his work, and something came up he couldn't answer, he'd call in his assistant from another room and he would come in to answer the question and then he would go out again. He was a person with a lot of charm. He didn't know much theory but he had good intuition about what sort of experiments to do. The experimental group was on one floor of the building and on the floor above was the theory group but I don't think the theory group paid any attention to what the people on the lower floor were doing. There wasn't any interaction at all as far as I could see., But that's very typical I think. People in a theoretical institute work on their own problems and don't pay much attention to what is going on in the other institutes. That wasn't true of Mott. I think Mott was always very involved in experiments. Mott worked with Frank in crystals growth and things of that sort. That was when Frank was doing work with crystals experimentally. Mott was much involved with it.

Hoddeson:

The flurry of development in Europe from about 1928 to 1932 or 1933 which resulted in the quantum theory of solids seems to have settled down by 1933 with the exception of the group around Mott in England where there was still quite a bit of work going on. Is that an oversimplified approximation?

Bardeen:

I think that's true. Just after the quantum theory was discovered they were applying it to anything they could think of applying it to, and the theory of solids is one of them. Then the neutron was discovered in 1932 and so nuclear physics became a very popular field and a lot of the theorists switched their attention. Only a few remained working on solid state problems. I think all of the leading theorists in that period from 1928 to 1932 at one time or another worked on problems which we now call solid state. In the late 30's, very few were working on solid state theory.

Hoddeson:

Mott was an exception. Let's see, is there anything else that was of interest about your trip to Europe?

Bardeen:

I found it very stimulating and picked up many new ideas. It was a very profitable trip. And I met people whose papers I had read and had admired but never met before. I guess through the prestige of the Bell Labs, they arranged so we could see their leading work. Physics was pretty much removed from Paris during the war because of the German occupation and Neel and his coworkers were at Grenoble, but he came up to Paris and met us there. We had a very fine reception at the Phillips Labs and met Mott. Also at Cambridge.

Hoddeson:

Was there any work going on in Scandinavia? I notice a reference to Benedicks in one of your papers of 1915 work on germanium and silicon.

Bardeen:

That was very early work. The first publication of work on germanium I think was by Benedicks.

Hoddeson:

I think there was a reference in your Bell lectures. [See p. 79 of J. Bardeen, "Semiconductor Research Leading to the Point Contact Transistor," Lex Prix Nobel en 1956 (Stockholm, 1957)].

Bardeen:

Not much was done until Lark-Horovitz picked it up during the war. He was the one mainly responsible for the activity in this country on germanium during the war. He never m and a useful rectifier for microwaves, silicon was better, but work they did with germanium was very important for our development.

Hoddeson:

When was the spreading resistance work done. Was that work done right after the war?

Bardeen:

That's Bray's work. I think that was just after the war. It was an outgrowth of the war-time activity.

Hoddeson:

Were you aware of the details of that study? By the time the war was over, I gather there wasn't quite as much information exchanged. Or was that not true? I got the impression that during the war, information was exchanged freely among the people who were working on common problems. Is that not true?

Bardeen:

Well, I wasn't involved in the activity during the war but after the war at meetings people talked. But of course travel was more difficult than it is now and we couldn't get together as frequently as we do these days.

Hoddeson:

I understand there was work going on at Sperry and RCA. Or was that relatively unimportant as compared to the work at Purdue?

Bardeen:

There were a lot of arguments in regard to patents for example with regard to doping. I've forgotten whose patent it was. Now who claimed credit for doping the 3-5 compounds. But I don't think his patent held up. They argued about it in the courts for a long time. I wasn't familiar with that work. I wasn't as familiar with that work as I don't think they published very much and we depended greatly on the reports which were published during the war, from the Radiation Lab, and from Purdue, and from the Pennsylvania groups. I think places like RCA and other places didn't publish very much and I wasn't familiar with what they did.

Hoddeson:

I think we're up to a long discussion on the transistor which starts with Shockley's design for a field effect amplifier. Shall we continue next time?

Bardeen:

I think that would be best.

Hoddeson:

Thank you again.