Nevill Mott

If I had been a professor in a different kind of laboratory (in Bristol) it would have gone differently. It’s funny to say that when we had Cecil Powell who later got a Nobel Prize, but at that time I don’t think Cecil was producing very much, whereas the solids group was. I think that one of form of physics as interesting another and that you should do what your date leads you to do.

[Question from interviewer] Bragg was at Manchester and my first job was in Manchester where I was a year, and he introduced me to scattering problems because electron diffraction was just coming in and one wanted the scatter amplitude from individual atoms. So he led to me writing Mott and Massey. [The theory of atomic collisions, 1933] Bragg splendid chap. But then they offered me a Fellowship here [Cambridge] and that was too good to miss. I must say I liked Manchester. So I’m afraid there’s no very romantic story there. How did I get interested in semi-conductors? Well we wrote Mott and Jones Properties of Metals then I had another colleague there, Ronald Gurney, and it was perhaps natural after getting a general understanding of metals to have a look at non-metals, and it was then I came across Pohl’s work. You know, of course, semiconductors were in evidence in my childhood, I remember trying to make a radio set work with a crystal detector. Also I believe, am I right, that copper oxide rectifiers with were in some kind of commercial use. [Confirmation from interviewer], very early on.

They were around and in my book with Gurney we tried to explain their mode of operation. Actually Schottky did it first but we hadn’t noted his work, and them we saw Pohl’s work; it was so absolutely beautiful. He published little paper after little paper setting out the experimental evidence never producing a theory — he didn’t believe in theories — but the experimental evidence so spoke themselves that the theory seemed to us obvious. When we started we didn’t know that an F-centre was a vacancy with an electron in it, in fact we felt it was a self-trapped electron what we now call a polaron; and I think I can remember Ronald Gurney coming along and saying that it isn’t that, isn’t a vacancy. We got all excited so we made contact with Pohl, [and] we to see him. I can’t remember, but we organised a conference on solid state physics in Bristol and invited him to come along with others. But Pohl I would call the father of solid state physics, I don’t know any earlier systematic investigation of anything in this area. [No knowledge of Baedeker — comment of interviewer] You know this book by Sommerfeld and Bethe [Elektronentheorie Der Metalle].

Bethe was an incredible man, and really he’d worked out everything, not everything that we can work out now with more sophisticated techniques, everything you can do with elementary quantum mechanics. I feel almost our Mott and Jones is almost a popular edition of Sommerfeld and Bethe, popular in the sense that it applies to real materials. Most of the ground work is in there. Mott and Gurney is, in a way, based on Pohl. I don’t mean all of it but a great deal of it. Then Gurney and I, why did we get on photography? Oh! again because of Pohl, the German group had been investigating — Oh! wait a moment, there were Frenkel defects a Schottky defects. Frenkel is another father of solid physics really isn’t he. He had invented the Frenkel defect the interstitial ion in the thing (?), somehow or other we had got.

Yes, in Bristol there was Professor Garner, Professor of Chemistry, interested in the photo-decomposition of azides. He was an explosive man, azides are detonators, he had blown two of his fingers off. Anyhow we were very friendly with him, you know what these chemists are, their pound up their azides they don’t take any trouble, this went off in his hand. We got interested in these problems and we thought, well isn’t the decomposition of silver bromide the simplest and most important so we cooked up a theory really based on Frenkel’s ideas in a way. Well there was the war, and I didn’t have any contact with solid state physics during the war — mainly on radar and operational research problems, but afterwards we came back and semiconductors had a future and industry was interested in them, and at Bristol we used to run summer schools for industrial visitors and tell them about Mott and Gurney and elementary quantum mechanics and what dislocations are. I remember some of the Dutch, the Philips people, coming over one of them saying that during the war we were, of course, cut off from books, we wrote Mott and Gurney for ourselves.

[Conversation concerning Mott and Gurney. Mott, ‘The band theory, as we now call it hardly started.) He [Pohl] used to publish a lot of it that little Göttingen home journal — I forget what it was called, but we established good relationships. [Question from interviewer about Born, Pohl had told the interviewer that the nucleus was the focus of interest]. Yes, but Born was, he’s another father of the solid state, [Question about Jammer’s writing, referring to The Conceptual Development of Quantum Mechanics] I think that once we had Pauli’s Exclusion Principle. No! I damned if I see how could done it without quantum mechanics. I think the semiconductor explosion both experimentally and theoretically is something I took little part in. The experimental work — there was singularly little of it in the universities. Industrial development, the Bell Labs and elsewhere, and theoretically the work was so tedious the whole thing became a matter of computing.

Once the techniques for exploiting band structure were developed these things could be worked out, the results were very interesting, but the detailed calculations were something that I would have hated to be involved in. [Question from interviewer about the involvement of industrial labs] Well, I think that they felt it was rather an unpromising field, for they would be competing with rather massive efforts particularly in America, and they wouldn’t have the technical facilities and so on to develop devices, the purpose of the work was to get the knowledge on which devices could be developed. [Question from interviewer nuclear physics more attractive] No, I don’t think that. there were plenty attractive of problems on solids. The early work on dislocations were all university based. And the work on superconductivity was mainly university based, this laboratory [Cavendish] until recently in solid state was almost all metals. It was either dislocations in metals or else the low temperature work, superconductivity, the Mössbauer effect and the shape of the Fermi surface.

One of the things that came out of Cambridge in the 50s and late 50s and 60s was the fact that the Fermi surface really a physical quantity whose form could be determined experimentally, and not just, as would appear from Sommerfeld and Bethe a mathematical fiction. [Question from interviewer — ‘Is this a difference of philosophical attitude] No, you start with a picture of non-interacting electrons, this gives a space separating occupied and non-occupied states. One’s first feeling, which I’m sure I had, was that if you put in the interaction between electrons it would all get ‘fuzzed’ out. Then this story that I mentioned showed that there was at any rate a sharp upper limit; and Pippard’s work in the mid-1950s and Landau’s theoretical work showed that the Fermi surface could be both measured and defined in a way that included interaction. That the real surface wasn’t fuzzed up by interaction. But semiconductors if I come to much more recent times, the Wilson separation of metals and insulators into those with full and empty bands and those with half full bands struck me that it couldn’t be right. Not universally right because of nickel oxide which had eight electrons out ten places in each nickel ion and yet it is a transparent insulator. This gave rise to an answer to the problem which has had the name of the Mott transition dubbed on it which I finally felt I had better write up into a book. The most interesting case us that of the heavily doped semiconductors, the lightly doped are extrinsic and the heavily doped become metallic.

[Here Mott moved into his 1974 interest, the performance of glass-like substances, this subject was not in the interviewer’s interest, so transcript was done.] Pohl, one of the neatest clearest experimenters who chose the ideal subject. Bethe, Smeckel, found after the war many were interested. [Mott did not feel inclined to evaluate his own work, or to compared his work with others, he continued] I’ve always had a suspicion of the over mathematical German school which isn’t only German now. Bethe is not in this category. [Comment from interviewer on pragmatic considerations]

After the war, the war was, of course an education for people brought up academically and we felt that we had something in understanding semiconductors and industry needed it. And it is indeed true that if I go to any lab one of middle management will tell be that they begun to understand semiconductors at one of our summer schools at Bristol, and found it very useful.