History Home | Book Catalog | International Catalog of Sources | Visual Archives | Contact Us

Oral History Transcript — Sir Rudolf Peierls

This transcript may not be quoted, reproduced or redistributed in whole or in part by any means except with the written permission of the American Institute of Physics.

This transcript is based on a tape-recorded interview deposited at the Center for History of Physics of the American Institute of Physics. The AIP's interviews have generally been transcribed from tape, edited by the interviewer for clarity, and then further edited by the interviewee. If this interview is important to you, you should consult earlier versions of the transcript or listen to the original tape. For many interviews, the AIP retains substantial files with further information about the interviewee and the interview itself. Please contact us for information about accessing these materials.

Please bear in mind that: 1) This material is a transcript of the spoken word rather than a literary product; 2) An interview must be read with the awareness that different people's memories about an event will often differ, and that memories can change with time for many reasons including subsequent experiences, interactions with others, and one's feelings about an event. Disclaimer: This transcript was scanned from a typescript, introducing occasional spelling errors. The original typescript is available.

Access form   |   Project support   |   How to cite   |   Print this page


See the catalog record for this interview and search for other interviews in our collection



Interview with Sir Rudolf Peierls
By Dr. Lillian Hoddeson and Gordon Baym
In Minneapolis, MN
May 20, 1977

open tab View abstract

Sir Rudolf Peierls; May 20, 1977

ABSTRACT: Student work with Arnold Sommerfeld, Wolfgang Pauli, Werner Heisenberg; first paper with Heisenberg; paper on Hall Effect; history of early contributions to the electron theory of metals; extrapolation of second Hall paper by Leon Brillouin to three-dimensional case; errors in Pauliís work on an-harmonic terms; Peter Debyeís error; paper on free electrons. History of and contributions to theory of semiconductors: Felix Bloch, Yakov Ilyich Frenkel, Ralph H. Fowler, Nevill Mott; Max Born on nonlinear electrodynamics; paper with Lev Landau on quantum electrodynamics; Walter Shockley. Search for superconductivity during 1930s. Knowledge of solid state physics in Germany and England beginning 1928; effects of World War II on development of solid state physics. Also prominently mentioned are: S. Chandrasekhar, Arthur Stanley Eddington; University of Cambridge.

Transcript

Hoddeson:

This is Lillian Hoddeson. Gordon Baym and I are interviewing Sir Rudolf Peierls during a Symposium on the History of Nuclear Physics in Minneapolis. Today is May 20th, 1977. Iíd like to begin this session by asking you to give me a little background on your working relationship to Pauli and Heisenberg in the late twenties and early thirties. I understand that you were working for both.

Peierls:

Well, yes, but not at the same time. I was a student under Sommerfeld, and in the spring of 1928 Sommerfeld went off on a sabbatical to the United States, I believe to Michigan but Iím not sure, so on his advice I joined Heisenberg, who had just started a theoretical physics department in Leipzig. I stayed for two semesters. Then Heisenberg went on sabbatical, and on his advice then I went to Pauli in Zurich. So I had my final semester with Pauli, I wrote my thesis there. And then I had to submit it to Leipzig, because I hadnít got the residency requirement in Zurich. So I submitted it while Heisenberg was still away, and my examiner was Hund.

Hoddeson:

How did you get started on your work in solids?

Peierls:

When I was in Leipzig, Felix Bloch had just written his thesis, on electrons in periodic potentials. So there was great interest in the theory of solids. Heisenberg, I think was then beginning on his work on ferromagnetism. After one abortive problem that Heisenberg suggested --

Hoddeson:

-- what was that on?

Peierls:

That was the following: Some experimentalist -- a German physicist called Rausch von Freudenberg had done some spectroscopy of Canal Rays, you know, the positive ions, and had found lines very much wider than he expected them to be. And Heisenberg suggested this might be a nice example of the Uncertainty Principle, because those ions could be visible in front of the spectroscope only for a short time. He wanted me to work that out. Well, being inexperienced, I did not work out the orders of magnitude on the back of an envelope, but I tried to set up a formalism. And of course, it wasnít obvious to me immediately that you should treat the motion of the ions as classical, which would have made it fairly simple. I set up a Schrodinger equation for the whole thing and then made approximations and so on, and in a few weeks time I had formalism. When I put numbers into it, I found that the broadening from that cause was I think three orders of magnitude less than what was observed. So I went back to Heisenberg and told him somewhat diffidently, ďI donít think this mechanism can do it.Ē Heisenberg checked my numbers and immediately agreed of course and said, ďOf course, how stupid -- the experimentalist must have given me some wrong numbers or I misunderstood.Ē So that was that. But it was good experience all the same. That was not in solid-state physics obviously. But, then, Heisenberg suggested that I should look at the problem of the wrong sign of the Hall Effect, which sometimes is positive and we understand why today, but at the time -- Well, it was clear that from the Bloch results you could sometimes get an acceleration, in the opposite sense to the wave vector. You had to convince yourself that this was real, and you had to convince yourself that that did not mean that there would be a negative conductivity, and so on. So that was the first paper that I ever published. [1] I enjoyed being able so early to make a real contribution to something.

Hoddeson:

Was Heisenberg in good touch with the experiments on the Hall Effect?

Peierls:

Not particularly on that. This was an established thing. You looked at any table of constants, and it was there. It was one of the -- you see, there was a long list of paradoxes, in solid state, in particular, in the electron theory of metals. One by one they then disappeared. So he looked at the next one and said, ďNow, how about that one?Ē It was a little bit harder to see.

Hoddeson:

The great start, I guess, was with the Pauli paramagnetism paper, then followed by Sommerfeldís extensions.

Peierls:

Yes, well, in the sequence, of course, Fermi statistics was the first step, and then Pauli applied it to paramagnetism. And then Sommerfeld, who knew much more about metals than Pauli.

Hoddeson:

-- having worked on them earlier, around the turn of the century --

Peierls:

-- right -- he spotted that this would account for the specific heat anomaly. Then he wrote a long paper working out all the details, many of which couldnít be handled by the model he had. For example, he treated the collisions as if they were classical kinetic theory collisions with obstacles and therefore he couldnít get the normal increase of the resistance with temperature and so on.

Baym:

This being electron-electron collisions?

Peierls:

No, electron-atom or ion. It was left to Bloch to show that a perfect lattice would have no resistance, and therefore some of the resistance was due to lattice vibrations.

Hoddeson:

Do you consider this little sequence of paper between Ď27 and around Ď30 or Ď31 to have been the really turning point in the beginning of modern solid state physics?

Peierls:

It certainly started it. Of course there was then a continuous development. Now, whether you can find any other landmark -- that is, not a complete break -- Iíd have to think. Certainly very many important things have happened. The point is you couldnít understand solids, without using the quantum theory, and that was the first time that the quantum mechanics was applied. That was the first step. I think the important thing was to have quantum theory or quantum mechanics. Once you had that, it was bound to come sooner or later. But I remember, one development which seemed for me very important, and that is, having written that paper about the Hall Effect, I was very conscious of the fact that this depended vitally on the behavior of the energy as a function of wave vector, which gives rise then to the reverse acceleration, you could prove that the average velocity of the electron, was just the group velocity, and therefore the derivative of the energy curve. That was all fine. But the shapes of these curves were clear only in the limit of tight binding which clearly wasnít a very good approximation for real metals. Nobody had thought about the other limit. What would happen there? And it seemed to me that first of all, in the other limit, with nearly free electrons, you wouldnít have that behavior, and that therefore this behavior could only be over part of the range. Then later, I think I was working in Zurich, it suddenly dawned on me how to handle the nearly free electrons, and that you get the breaks in the energy curve there. And that even in first order, a weak field you already had qualitatively the Bloch picture. I still remember the excitement. What you have to do is simply to solve a pair of simultaneous equations. Those are the same equations which were familiar from the anomalous Paschen-Back effect -- itís always the same thing when you have two effects which are comparable.

Baym:

These are the equations that are in all the textbooks now?

Peierls:

Yes. I remember how thrilled I was to see that. But I only did it for the one dimensional case, because it seemed to me enough to make the point, and then Brillouin picked that up, and he quoted my paper, and he discussed it completely for the three dimensional case, and got what is now called the Brillouin Zones.

Hoddeson:

Do you remember how you got to those equations?

Peierls:

Well, I donít remember it in detail, but I suppose in the way in which theyíre now presented in textbooks. In general, if the potential is weak, the perturbation theory is all right. But when you are exactly at the edge of what the textbooks now call the Brillouin Zone, then you have two states which are degenerate, and then you should make linear combinations. And thatís easy. So far thatís obvious. But if the energy difference is comparable with the potential matrix element, then you have to still take a linear combination but retain the energy difference when the matrix element of the perturbation is comparable to the energy differences you have to do that.

Hoddeson:

Did you then go right away and show it to Pauli?

Peierls:

I donít remember. The point is I then made the mistake of putting in too many things into the same paper. [2]

Hoddeson:

This was your thesis?

Peierls:

No, this was afterwards, because the thesis was about insulators not metals. This was next. Perhaps I can tell you about the rest of that paper a little later, because it follows after my thesis. Anyway, this was really hidden away in a paper that contained all sorts of different ideas, so not many people noticed the point. There were a couple of small papers in Leipzig, which -- one, about discussing the variation of the ionization energy in what we now call the P shell, of the elements in the first row of the periodic table. Well, I was irritated -- thatís how one often gets onto interesting problems -- I was irritated by a paper by somebody who discussed this behavior which looks like a closed shell in the middle of a row in the Periodic Table.

Hoddeson:

Do you remember who it was --

Peierls:

No, I donít. But I think I quote him in that paper, so I can look it up. Anyway, that wasnít very deep. But then, I also wrote that paper, proving that in one dimension any potential, well, however shallow, will have at least one bound state, which Wigner mentioned today. And to which I then added the wrong conclusion at the end. But then I went to Zurich. One thing that now looking back I find very hard to believe, there must have been more hours in the day at that time. Because I couldnít believe the dates, until I checked them. The time scale was as follows. I moved to Zurich in time for the summer semester, which started in April, the beginning of April. I found myself lodgings. I contacted Pauli and organized myself in the department, translated a book by de Broglie into German, which Iíd promised the publisher to do. They were in a very great hurry, because the book wasnít published in French. No French publisher would take that risk. It was published later in France. But it was to be published in English and German translation, and obviously whoever came out first would sell most of the copies, and so, the publisher was pressing me to do it in a hurry, and I did. Well, then, Pauli suggested the problem for my thesis. I sat down and solved the problem. I wrote up my thesis. I typed it myself. I submitted it to Leipzig, because I couldnít submit it in Zurich. Then I think I had my oral examination for the Ph.D. on the 23rd of July, but Iím not sure about that date. It might have been August.

Hoddeson:

Was Pauli doing any work in solids at that time? I thought he got out of it himself.

Peierls:

He got out of it, except for one problem, which was not really quantum mechanics. He had always been interested in the sort of Born-von Karman theory of specific heat and vibration of a linear chain. He was interested in the effect of an-harmonic terms in that theory. And I think probably the motivation was because that was a good model for field theory. I donít know. There Iím guessing. He didnít say that. But the point is he had done some calculations about this, which should give the damping of sound waves, which is due to anharmonicity. On this problem heíd done a one dimensional calculation, and heíd presented the result at a conference. It was never published but the abstract was published of the talk he gave to a meeting of the German Physical Society, and what he said there was completely wrong. The only wrong statement I know of published under Pauliís name.

Hoddeson:

What did he say?

Peierls:

Well, he somehow got damping due to, three phonon transitions, as we now call it, one phonon going into two and two into one, in a one dimensional chain and there arenít any. Now, of course, I didnít see the details of that, but he showed me a couple of pages of notes relevant to this, and I think what was happening was that he took the simple Born-von Karman case where the frequency is essentially in KA/2 and he then wrote down the conditions for the conservation of wave vector and energy, and got certain solutions. I think his solutions meant that one of the phonons had the initial K plus 2Π/a or something. In other words, there was no change, but he didnít notice it, because he expressed it in different variables. I think that must have happened -- I donít otherwise remember the couple of pages sufficiently to say. But I mean, he also knew, there was some doubt he obviously suspected what he had done, and thatís why he told me to look at it. Well, then, I found, to my great amazement, that everybody who had worked on that subject was wrong. Debye had a persuasive theory, which gave probably the right temperature dependence, but was quite wrong in concept. I donít know whether you know about what is wrong there? Debye said, very nicely, ďAnharmonistic simply means that the refractive index, in this case the sound velocityĒ -- the refractive index for the sound waves, depends on density, and therefore, if you have density fluctuations, you get scattering. And therefore, we can simply take over the known theory of scattering of light, by the density/fluctuations caused by the sound waves. Very elegant. What he left out of consideration is that this theory assumes that the density fluctuations are stationary. Now, for scattering of light, thatís fine, because compared to the light velocity they really donít move. But for the scattering of sound waves itís fatal, because the density fluctuations move at precisely the same velocity as the sound waves, because theyíre the same thing. Therefore he gets the dynamics all wrong. In particular -- his model would give damping in one dimension, where there arenít any such things. It also would get a finite conductivity for a continuous medium, where there are no Umklapp processes and as we now know, there is no resistance.

Baym:

Variations of this have been used by Landau and coworkers for liquid helium.

Hoddeson:

Recently?

Baym:

In Ď50ís.

Peierls:

Yes, right. And then there was a question whether they give a finite thermal conductivity or not. They didnít fall into that trap. I donít think they used the Debye model. I think they knew better. So, that led to the idea of Umklapp processes, and therefore, to the prediction that the thermal conductivity of a pure crystal should rise exponentially at low temperatures. It took about 20 years before that was verified experimentally. And the reasons why it took so long are interesting. One is that I expected, and certainly made it sound as if it should be obvious, that it should happen at much higher temperatures. The reason was, one normally talks sloppily, about high temperatures and low temperatures, -- meaning high temperatures where things are completely classical, and low temperatures where things are in the extreme quantum limit and all wavelengths are long and so on. And the measure of that is related to the Debye temperature. But the Debye temperature, in its definition has an inconvenient factor of 2Π or something, as a result of which the Debye temperature (N) is that at which youíre practically completely classical. So to say that above (unknown symbol) is a high temperature is fine. But to say that below (unknown symbol) is a low temperature is not right. You have to go very much lower. But the other thing is much more disgraceful, namely, I didnít notice (or didnít stress) that ďpureĒ meant in this case, also isotopically pure because an isotopic mixture, a random mixture would of course, be irregular from the dynamic point of view, because the motion of the atoms, depends on their mass, and therefore you would get scattering.

Now, Pomeranchuk wrote a paper, in the 1940ís, pointing this out but in wartime, the Journals didnít arrive or people were too busy to read it, so it wasnít noticed. And it wasnít until Berman at Oxford did the experiments, and then noticed empirically that some substances showed this exponential rise, and others didnít. He then got on to the fact that those which showed it were isotopically pure, or nearly pure. Then of course he knew what was going on. Another interesting thing is, I then wrote a paper -- now Iím coming back to this paper which also contained these arguments about the nearly free electrons Ė- I pointed out that you had the same problem in principle with metals, because in metals, unless you have a Fermi surface which touches the borders of the Brillouin zone, and so on, you also had problems in getting Umklapp processes and disposing of the pseudo momentum. And therefore, in principle, pure metals ought also to show, in most cases, this exponential rise. That nobody believed. It was verified only a few years ago. It took a bit longer. But the reason why that wasnít spotted was it is very hard to get metals really that pure to see the ideal resistance at these low temperatures. What is normally done is to take Mattheissenís rule for granted, so that you say the impurity resistance and the ideal resistance are just additives and therefore, you deduce the ideal thermal resistance by subtraction. But if you have enough impurities, then of course you neednít worry about disposing of the pseudo momentum, and so you donít see that. Of course that is what is now described as phonon drag. Itís the same thing.

Hoddeson:

I wonder if you could comment on this remark that Wilson made, in another interview, that this theory of metals is really a reformulation of certain ideas that he found in your papers.

Peierls:

Well, I think that Wilson is at heart a mathematician out of the Cambridge school, and heís more concerned about mathematical formulations that I would be. Now, physically, I think he has certainly made one important contribution, and that is to discuss semiconductors seriously. He initiated the ideas about semiconductors. Of course, he didnít finish it. There was very much more to be done. But he started it. Now, Iíve heard, there is somebody who is studying the history of solid-state physic also, somebody in England, Iíve forgot his name, who has an interview with Wilson that you may have seen. And according to that, Wilson also originated the idea that a metal is a case with a partly filled band. While in an insulator all bands are full or empty. Now, this certainly surprises me, but I cannot contradict it on definite evidence. I mean, in my paper on the Hall Effect -- not in the main paper, but in a short paper in the Physikalische Zeitscrift, which is a summary of a talk I gave at some meeting -- has actually a diagram which looks like this: (figure drawn in the transcript) of an energy surface, which almost fills the band and where then I said, ďThis is a case where you obviously get the positive Hall Effect.Ē Now, I didnít actually say that if you fill also the last holes here, you get an insulator; seems to me, looking back, however that that idea must have been quite obvious. But you know, the discussions this afternoon[3] showed, how many things were not obvious at the time Ė- I certainly never wrote down what was the difference between metal and insulators. But I am really fairly clear that Bloch must have known that, although Wilson claims that Felix Bloch was surprised at it.

Hoddeson:

And that he wouldnít believe it, yes.

Peierls:

Now, I think itís certainly important to find out what Felix remembers. [4]

Hoddeson:

Yes. Well, sometime I or someone else will certainly ask him about that.

Baym:

Did you come to these ideas yourself also -- such as having filled bands in an insulator.

Peierls:

Well, I think so. But I canít prove it. I never said it in writing. You see, one must be very careful, in talking about what I think I knew at some particular time, because so often you discover that things which you thought were obvious from the beginning, in fact, originated some years later.

Baym:

About when would you say that you did understand this?

Peierls:

Well, my guess would be, it was likely that I knew that when I wrote the paper about the Hall Effect. But I canít be absolutely sure. If I didnít know it at that time, then I canít say when I knew it. I know I know it now.

Hoddeson:

Bloch was working at the same time in Leipzig as you were working in Zurich, is that right?

Peierls:

Well, he was in Leipzig when I was there. Then, when I went to Zurich, I think he stayed in Leipzig, but he came frequently to visit Zurich because it was his home town.

Hoddeson:

How frequently?

Peierls:

Oh, I donít know -- a few times. During that first year, he was probably there a few times. Certainly, he was there in the summer of Ď29 when I had only just got there. There was a nuclear physics conference in Zurich, and he came to that; and it was a nice occasion to meet with many other people.

Hoddeson:

And did you discuss the theory of metals with him then?

Peierls:

No doubt. I no doubt told him what I had done, and he no doubt told me what he was interested in.

Hoddeson:

Were you aware of other people who were working in this particular area at that time? Or were these two centers -- Leipzig and Zurich -- the main centers?

Peierls:

Well Ė-

Hoddeson:

-- I understand there were some Russians working there too occasionally? Frenkel.

Peierls:

Frenkel was very active. And now, at what time -- one always forgets about the relations and sequences of the times, but --

Hoddeson:

-- Had you read Frenkelís papers?

Peierls:

Oh yes. Yes. In those days you read everything that was published in quantum mechanics. It wasnít so much. I mean, quite soon --

Hoddeson:

We were on Frenkel, and communication with others.

Peierls:

Yes. Well, very soon, of course Brillouin got into the act. Then I heard about Wilson, and that started out badly because he wrote a paper in which he said, the whole theory of Bloch and myself was wrong, because it used cyclic boundary conditions, and these were obviously wrong -- in solids -- and certainly you should expand things in standing waves, and then, various results came out differently. So, I decided to reply, and in fact, I replied first of all, since Wilsonís paper was, I think, in the Royal Society Proceedings. I thought maybe my argument should go there. So I wrote to Fowler but I didnít write the paper first, I wrote to Fowler, whom I met. Would he communicate a short paper disagreeing with Wilson? He replied that he certainly would do that, provided the paper was reasonably written, and he said, ďOf course, please remember that in a paper, you can call somebody a fool, but you shouldnít call him a damn fool.Ē (Laughter) However, this letter was delayed because Fowler wasnít in Cambridge and by that time, having not heard from Fowler, I sent it to Zeitschrift fur Physik. That had a rather interesting and amusing sequel. Fowler had the habit of going to sleep at seminars, but he had trained himself not to open his eyes when he woke up, to continue having his eyes closed, and then open them and ask a relevant question, so he gave the impression that he was listening all the time. One day -- I wasnít there, but Iím told -- Brillouin was visiting in Cambridge, giving a talk, Fowler went to sleep as usual, and there was something in Brillouinís talk in which he had to mention this controversy, with Wilson sitting there. Brillouin being French and very polite, had thought of a very careful wording. He gave his reaction to that situation, which basically agreed with me, but he managed to say that in a form which wasnít too offensive. Then Fowler woke up, there was a discussion, and when it finished Fowler said, ďWell, by the way, you must know, how is it now with this disagreement between Peierls and Wilson?

Baym:

Did you read Fowlerís early papers on applying Fermi statistics to the matter in White Dwarfs? [5]

Peierls:

No. I heard about it, and I was interested in it. I got into that a little later, because Eddington, as you know, didnít believe that you should use relativity for that. Remember that Chandrasekhar extended the formula into the relativistic cases. Eddington disagreed, not because of any data or anything, but because he liked the non-relativistic formula, and thought it ought to be right. Then he wrote a paper saying that relativity is wrong, when applied to that situation. You couldnít combine it with quantum mechanics. When some people pointed out that after all, the Dirac equation for the hydrogen atom worked pretty well. He said, ďWell, it was a question of symmetry. In the spherical geometry itís all right. But in deriving the equation of state, you usually divide the phase space up into rectangular boxes and in order to count the density of states thinking classical. It seems to me a star is more like a spherical box than a square one. (Laughter) But then, Pryce, Dirac and I together wrote a paper refuting Eddington argument, and proving that the boundary conditions didnít matter and the shape didnít matter, and you could derive the states any way you liked. Thatís my only contact with that subject.

Hoddeson:

Going back to the earlier period --

Peierls:

-- yes --

Hoddeson:

Did you have any interactions at that time with Mott? His work was a little bit later.

Peierls:

Very much later. I mean, he -- I think -- was interested in relativistic electron scattering, and interactions, things like that, Mott scattering. He got interested in solids, probably only when he moved to Bristol, which must have been about Ď34 or Ď35.

Hoddeson:

You were by then no longer actively working on solids.

Peierls:

I was still interested, but I was not in direct contact with Mott. In Ď33-Ď35 I was in Manchester and Mott was in Bristol. We did meet occasionally, and corresponded about some things and --

Hoddeson:

Is the correspondence existent?

Peierls:

I donít have any. I donít think I have any copies of that. Mott may have something. In fact, recently, in a talk, he gave me credit for some suggestions which Iíd made on problems he was working on which Iíd completely forgotten, so there was some interaction.

Hoddeson:

Do you remember what the subjects were that you were discussing?

Peierls:

No. No. I mean, I am now trying to remember what the problem was which Mott said produced the important argument.

Baym:

When did you begin work on what is now known as ďPeierls Instability?Ē

Peierls:

Oh, that was when I was writing my book on the quantum theory of solids, [6] in -- well, that book was published in Ď58, I think -- so that must have been Ď53. Well, a bit later, in Ď53 I was lecturing, teaching at the summer school in Les Houches and the book was based on those lectures, so it must have been early Ď54.

Hoddeson:

Again going back to the thirties Ė-

Peierls:

OK.

Hoddeson:

Iím interested in your interactions with Born, if any. He was interested in solids a little bit.

Peierls:

Yes. Well, of course, I used his lattice theory a lot. But I didnít have any contact with him before he came to Cambridge and after he was in Cambridge, I saw a good deal of him both in the summer of Ď33 and then Ď35 to Ď37 when I was again in Cambridge. But then he was not working the theory of solids. He was working on his nonlinear electrodynamics, things like that.

Hoddeson:

And so your discussions tended not to be on solids?

Peierls:

Yes. I think the answer is simply that we didnít talk much about physics. We were good friends, but I donít think we talked about what either of us was working on. [7]

Hoddeson:

How about Landau? Did you know him?

Peierls:

Oh yes, a very interesting person, very impressive. I met him first when he came to Zurich in Ď29, he could only stay a short time, because then Switzerland and the Soviet Union had no diplomatic relations. It was a special concession that he was allowed to come, in particular having a scholarship from the Soviet government, and he was allowed to stay in Zurich for, I donít know a month or six weeks, and then the permit was extended for a while, and then again for a shorter time, and so on. Landau was very flattered because he said, ďLenin stayed in Switzerland for three years, and didnít managed to start a revolution, but they obviously think I can.Ē (Laughter) The next year, he was back with a Rockefeller fellowship. Then, if Mr. Rockefeller would give him money, then it was O.K. to stay for a year.

Hoddeson:

He was working on diamagnetism.

Peierls:

That was part of it, yes. He came to Zurich with this theory of diamagnetism essentially complete. However, he immediately saw the point of that and was very impressed and liked it, discussed it. Then Landau and I got involved in all sorts of things, for example, we wrote a paper then on quantum electrodynamics in configuration space, using coordinates of the photons as variables which turned out to be mathematically cumbersome and physically very unreasonable. We were interested simply in establishing that this was a possible way of describing things.

Hoddeson:

What was the nature of your working relationship with Landau?

Peierls:

Oh, we discussed things a lot. We got together a lot. I think while he was in Zurich, we spent a very large amount of time together.

Baym:

Did you speak German together?

Peierls:

Yes. I canít remember all the things, we discussed, but certainly he was then already very interested in astrophysics and so on. We discussed that, but I donít think I contributed much. He obviously was a very impressive character. I mean, his way of working was -- if he saw a paper that interested him, he would glance at it to see what the problem was, and as a result he would sit down and solve the problem himself, and see whether his answer agreed with the paper.

Baym:

While weíre on the subject of Landau, just tell for the record, the story of Landauís remark on nuclear energy?

Peierls:

Oh yes. That was in 1934, during the summer when I was on a walking trip with Landau and a friend of his, and one day I said, ďHow is that about nuclear energy? Is that just science fiction, or is there something real?Ē And Landau answered immediately. ďWell, you see it is difficult, because of course there are nuclear reactions which release energy, but they have to be started by charged particles bombardment, and most of the charged particles will lose their energy in passing through matter, and so you lose most of the energy. Neutrons are better, because they donít waste their energy. But the only way we know to make neutrons today is by charged particles bombardment, so weíre back to the same problem. But if somebody discovers a reaction which is initiated by neutrons, and produces secondary neutrons, then youíre all set. This was about two years after the neutron was discovered.

Baym:

Wonderful.

Hoddeson:

Did Landau write anything about that?

Peierls:

No. It was an obvious thought to him and was not interesting unless someone really discovered a suitable nuclear reaction.

Hoddeson:

How about superconductivity? Was that much discussed that period?

Peierls:

Oh yes of course. You see, since it was a period when all the unsolved problems suddenly found a solution, it was obvious that one must find out what is going there on in superconductivity. And therefore, anybody who had a new idea relevant solid state, to metals, immediately tried to see whether this could explain superconductivity. Very often, one rushed and said, ďAh of course this might do it.Ē I mean, for example, when I saw that conservation of pseudo-momentum and lack of Umklapp processes might sometimes lead to a very high conductivity, rising exponentially, I immediately said: ďWell, now, could that be superconductivity?Ē It clearly wasnít, because I would (???) fast enough and was too sensitive to impurities. Thatís when Bloch announced his famous Bloch theorem, that ďtheories of superconductivity can be disproved.Ē (Laughter)

Hoddeson:

I didnít know about that. Letís see, was Heisenberg also working on superconductivity then?

Peierls:

Not then.

Hoddeson:

Pauli?

Peierls:

I think Heisenberg took it up after the war. Pauli, no.

Hoddeson:

You worked on it a little bit?

Peierls:

Only to the extent that everybody else was, in the sense that we were aware of the fact that here was an unexplained phenomenon and any new point of view would be immediately examined to see; could this be connected with superconductivity? And usually it wasnít. The people who were actively trying to make a theory of superconductivity at that point were Londons.

Hoddeson:

Were you in touch with their work?

Peierls:

I read about it, and was very suspicious of it, at the time probably too much so.

Hoddeson:

The Meissner Effect of course came out just in that period when you were working in 1933.

Peierls:

Yes.

Hoddeson:

Was that discussed much in the community?

Peierls:

It was just adding to the mysteries. Well, I was involved in a little bit of theoretical work, only in about Ď35 or Ď36, when I was in Cambridge -- about the one form of the Meissner Effect, where you get the intermediate state. If you put, for example, a superconducting sphere in a magnetic field and then I managed to get a rather nice phenomenological description of what was going on, later and somebody gave a proper microscopic theory.

Hoddeson:

You left the theory of solids around the mid-thirties?

Peierls:

Yes.

Hoddeson:

Did you become more interested in nuclear physics, or was there some other specific reason?

Peierls:

No, in fact, I always like to keep my finger in many pies. I never made a deliberate decision not to continue doing solid state physics, and Iíve come back to it from time to time. But one looks around, and if you see an interesting problem, you try to do something about it, if you can see you can do something.

Hoddeson:

How did the fact that you had to leave Germany in Ď33 and move to England affect the choice of the problems that you worked on?

Peierls:

Not very much. I mean, obviously, in your choice of problems, you are influenced by your environment, by who you talk to, what you hear about and so on. But the change was just as much from moving around within England, as from one country to another.

Hoddeson:

In fact, the center for solid state physics, as you moved from Germany to England, moved also from Germany to England, more or less, as it happened.

Peierls:

Yes. Well, there was a growth suddenly at that time, a growth of solid state physics in England.

Hoddeson:

Yes.

Peierls:

Not only some people arrived, who were interested, but also Mott became interested, Wilson was already there, and -- well they were not very big schools. England was peculiar in the sense that pre-war really the only center was Cambridge. And I know that for example Mott in Bristol was very cagy about accepting graduate students, for a Ph.D., because he said, ďTheyíre probably not as good as the people in Cambridge and what will they do? Was there a career for them? That only changed later. So he never had in that sense a ďschool.Ē He had a few students. Nabarro was one of them. Fuchs was another.

Hoddeson:

Did you teach solid state theory in England? Iím interested in how the new theory was propagated to students.

Peierls:

No. You see, in those days, it was not normal even to teach quantum theory to undergraduates. And there were no lectures normally for graduate students. The number of graduate students in any one place was too small to run any definite lecture courses. I remember, you gave occasionally a series of lectures for graduate students. I remember giving in Cambridge a course of lectures on nuclear physics. That must have been in the Ď35 to Ď37 period. In the pre-war period, I donít remember giving a course on solid state.

Baym:

What contact did you have with people doing solid state physics in the United States in that period? Wigner, Seitz, Bardeen, Slater?

Peierls:

None. One read their papers of course. Wigner was occasionally visiting Europe. He was in Manchester for quite some time in Ď33. In fact, he told me that this was because he was dismissed by Princeton University at that time because they thought he wasnít a good enough physicist. (Laughter) But that was before he got interested in Wigner-Seitz model. No, one met people at meetings but there werenít so many international conferences then.

Hoddeson:

One question about the phrase ďsolid state.Ē Do you remember when people first began using that term?

Peierls:

No, I donít remember.

Hoddeson:

Do you recall anybody using it before the war?

Peierls:

No. It may well have been. I mean, my guess would be, probably, yes. But --

Baym:

When did solid state physics become thought of as a separate discipline?

Peierls:

Gradually. I mean, I donít think one can pinpoint the time.

Baym:

You donít think that the war played any role?

Peierls:

Well, I think, at about the time of the war, during the war, of course, physics grew to the point where it split into sections, working on different things, and you were not just a theoretical physicist, for example, but you were a solid state theorist, a nuclear theorist, and so on. Before the war that didnít exist. So the whole idea of separate disciplines Ė- Now itís true, experimentally, that was always different. One couldnít at the same time work with nuclei, even if you only used natural wavelength sources and with low temperature things work. But it was much more the feeling of one subject, even if you were doing different things in that subject. And I think it was just the larger scale which developed that led to the greater specialization. In connection with that, you then thought of one area, as solid state.

Hoddeson:

In just looking at the number of papers with new solid state concepts in them, there is a peaking in the period Ď28 to about Ď33, and then it drops. I mean, there are still lots of solid state papers, but fewer with major new ideas -- until after the war. I always have made the assumption that the war played an important role by providing the technology for producing better samples, so that the theoretical ideas could be tested. Was that a factor?

Peierls:

That is probably a factor. Now, of course, just the number of physicists increased after the war, because of greater interest in science, greater support for science, and for physics. Then, also, there had been a delay -- people had neglected problems because they were all on war work. And they came back and picked up the ideas, after.

Hoddeson:

Yes, but solid state theory slowed down even before the war, in the late thirties. That is very basic work, such as the fundamental work that came out in the period immediately following the application of the Fermi statistics Ė-

Peierls:

Ah yes. Well, I mean, that was natural. There you had a situation where there was a breakthrough, and you exploited that, and when it was finished, you had explained all the outstanding paradoxes and seen that things work in general -- and then it became less exciting. To do new things which were not just routine applications became harder, as it does in every subject.

Hoddeson:

I wonder if we could perhaps end this session, which has been very interesting, on a more general note, on the effect of the war on solid-state physics in general? Weíve already discussed one or two of the effects. It seems there were so many techniques that evolved out of the wartime needs, like the microwave techniques.

Peierls:

Right, and then, generally, more materials, more funds for research, salaries, a greater scale of everything. I think also the expectation that solid state physics would sooner or later produce something tangible for industry, it wasnít quite clear where that would be, except, in the sense of materials, I mean, the connection between the electron theory of metals and metallurgy was always obvious.

Baym:

When did you become aware of the possibility of transistor action in semi-conductors?

Peierls:

When there was all the publicity about it -- certainly not before. That there were interesting phenomena in surface layers, I knew Shockley in Germany, was very interested in boundary layers and I had a long correspondence with him, trying to work out various hypotheses. I think he assumed certain approximations and certain conditions and I thought maybe one could do it otherwise, and I wrote down equations. I could never persuade Shockley that I was right and I never was persistent enough to carry these calculations out completely and publish them and so on. I was interested, but I didnít know enough about the problems of electronics to see how one could use the properties of these layers?

Hoddeson:

Well, I think weíve covered all the points I had questions about. We could go on some other time perhaps.

Peierls:

OK.

Hoddeson:

Thank you very much.

[1]2 Physik 53, 1929, 255-266.

[2]Ande Physik, 4, 121-148.

[3]Conference on the History of Nuclear Physics in Minneapolis, May 1977.

[4]Note added Jan. 1979. I have since consulted Felix Bloch, and his recollection agrees with Wilson's.

[5]R. H. Fowler, "Dense Matter," Roy. Astr. Soc. Monthly Note 87, 1926 pp. 114-122.

[6]P. E. Peierls, "Quantum Theory of Solids," (Oxford, London, 1955).

[7]Note added Jan. 1979. This was an oversimplification. Remembers talking with Born about this non-linear electrodynamics but not about solid-states problems.