Sidney Millman - Session II

Hoddeson:

The last time I spoke with you, you told me about your route to and through the Bell Labs: through the microwave work at Columbia, Jim Fisk’s recruiting of you, your work in the Electronics and Television Department with McRae.

Millman:

Radio and TV, yes, OK.

Hoddeson:

And then your work with W. H. Dougherty on traveling wave tubes.

Millman:

Well, work with Dougherty isn’t quite right. I mean, I worked with traveling wave tubes and Dougherty was the director of that particular department at that time. In other words, I merely pointed out that I was not in physical research at that time. Then in ‘52 I made an abrupt change, and I became Director of Physical Research. I didn’t start within physical research and get promoted to a higher position, but it was a complete shift, from things like traveling wave tubes, obviously applied physics, to a laboratory or department that deals principally with basic research, the physical research laboratory.

Hoddeson:

Were you happy to get back into physics research?

Millman:

Well, now, let’s put it this way –- yes. Certainly my arm wasn’t twisted in any way. It looked like a natural sort of thing. No, I think the most important thing was, not, back to physical research, but essentially whether the idea of getting into a position of administrative responsibility where there would no longer be much opportunity to do your own research — in other words, was this a good thing? And I sort of realized that as you get older, probably it’s a good thing. But I mean, one shouldn’t say that the chief attraction was the fact that, here, I am back in pure physics, although it’s certainly a very nice place to go back to.

Hoddeson:

Well, can you tell me a little more about it? It was being able to be in an administrative position where you could guide a lot of physicists in directions…

Millman:

Yes, let’s put it this way — if it had been a different position, director of some other laboratory, I probably wouldn’t have turned it down either. I mean, it’s sort of a stage in a career, and having had some experience with administration, as you do at the department head level, where you do partly your own research and partly administrating, and having been flattered, or given some indication that you’re doing a reasonably good job, you didn’t resist. It seemed like a good thing to do, and it wasn’t only because it was about physics that was fine, but it wasn’t only because of that, I don’t want to give you the impression that if it hadn’t been for the physical research laboratory, I would have not taken such a promotion.

Hoddeson:

I understand.

Millman:

OK, fine.

Hoddeson:

Now, I’ve pulled out a chart for July 1952, and I was wondering — we started doing this last time but hadn’t gotten into very much detail — if you might tell me a little bit about what each group generally was focusing on.

Millman:

Well, I should say, yes, it’s easy enough. First of all, the most important change around here was that this particular group under Grisdale — applied physics of solids — was brought in from the development department. They were not in physical research prior to January 1, 1952. The other parts of physical research remained more or less the same with some small changes. You see, Shockley, Morgan, and the many familiar people in each of these departments were all very much part of physics before. But this new group was in applied physics.

Hoddeson:

I see people who — well, certainly Geschwind is still here. What were they doing in applied physics?

Millman:

Well, some were involved in magnetism research or in crystal growing, but the chief sustaining activities in that department centered around Hogan, and Hogan was the leader at that time in the research on the gyrator, this microwave gyrator, using ferrites that you can use in circulators and isolators. You know about this gyrator? It’s an interesting device. The ferrites, and later on the garnets, have the interesting properties that they are magnetic but not metals. They have two sublattices. Some of the magnetic moments point in one direction, others in the opposite direction, and they’re not exactly balanced. At least you can arrange the composition so they are not — this is true of the ferrites as well as the garnets. Some of the iron ions are double plus and some triple plus. At any rate, the point is, they have some magnetic moments pointing in one direction, some in the other direction. These are atomic moments, we’re not talking about nuclear moments. The net effect is that you have an unbalance in magnetism, and that unbalance is appreciable, maybe say 10 per cent of what you’re getting in saturated iron or nickel, or perhaps even 20 or 25 per cent. It’s appreciable. But since these are oxides, they can transmit electromagnetic waves. Metals, of course, don’t, except for skin effects. So therefore you can put them in, in a waveguide carrying microwaves, and if you apply a magnetic field by Faraday rotation and so on, you can get some non-reciprocal transmission. The generic name was the gyrator. So you can imagine arranging things so that electromagnetic waves are propagated in one direction, and not back. Non-reciprocal. And therefore, it’s very much what was used, and this is very useful in microwave devices. For example, if you use a radar setup, you’d like to have the transmitted pulse going out, and not reflected. The much weaker return signal goes through a parallel channel also unattenuated. Well, this one-way transmission is very important — it helps give you isolators, for example, you can isolate one channel from another. They have circulators. You can direct something to go one way and not come back. This is a technical aspect of it. Research on the gyrator was the central activity of this group.

Hoddeson:

Did he invent that at Bell Labs?

Millman:

Yes, he was at Bell Labs, yes. There was some early history about somebody by the name of Telegan at Philips who had similar ideas. Whether he was the completely new inventor or not, that’s a little detail, but anyway, the fact is that he pushed it, and that became a very important application in microwave transmission.

Hoddeson:

Was it used then in your group for research?

Millman:

Well, this gyrator became a very important business for microwave propagation, and the activity in using it in microwave systems was actually done in Holmdel where they were interested in the propagation of waves through a circular pipe. I remember there were frequent meetings between people at Holmdel and Hogan and some of us from Murray Hill used to go beck and forth. They were the users, Gardner, Fox and others. They were the users of possible systems, whereas Hogan was interested in the physics of the propagation, how it works and what materials are best and so on, you see.

Hoddeson:

Why was the applied group added to the physical research group? My impression was that since 1945 or so there was a feeling that there should be separate pure research.

Millman:

That wasn’t diminished, except for the fact that this man Hogan and Grisdale were then technical leaders of an applied research group in the development area. However, M. J. Kelly, the Executive Vice President of Bell Laboratories, felt that the physical research laboratory would be a better home for this group. It was a little too researchy for the development area. In those days the development people tended to confine their activities to problems of design for manufacture at Western Electric Company, and this gyrator was sort of a very novel idea and it looked as if you’d have to understand a lot more of the physics of it. So they put it into physical research, but the main emphasis was its application to communication. They therefore called the department applied research.

Hoddeson:

Actually the transistor work, compared to the emphasis in the other groups, was also more applied by now, wasn’t it?

Millman:

Well, some phases of the transistor work was applied research. Returning to Grisdale’s department, there was also Geschwind who wasn’t interested much in application, but he was relatively new. I recruited him from Columbia, and he was interested more in pure magnetism, magnetic properties of sublattices, and what happens as you vary the temperature, and what accounts for the line width, but all of these were interesting for applications, because of the line width, losses and so on — there was the impetus to have people interested, not only in the idea of the gyrator, people interested in pure magnetism, so as to understand the mechanism of losses. For example, one of the things that distinguishes garnets — I don’t know if you know much about magnetic garnets — the interesting thing about the garnets is that in the garnets, all of the metal ions are trivalent, the oxygen is the usual double valence, and it leads to much less loss, whereas the ferrites have nothing but iron metal ions, some bivalent, some trivalent, and therefore, there are losses by the hopping process, electrons going from one to the other, trivalent, divalent, therefore giving you a mechanism for loss. Although in those days, it was only the ferrites. The garnets weren’t discovered for several years later. That comes later. But anyway, the ferrites at that time had the interesting thing about having great potential applications. At the same time, in order to understand better, for example, what to substitute for iron, to make it better at higher frequencies or lower frequencies, how would you affect it? So that was a big business at that time, and it was a function of that group, at least some people in that group, and Geschwind was one of them, to explore the physics of the magnetic properties. What happens when you have two sublattices? What happens if you vary the temperature? What accounts for the losses?

Hoddeson:

I understand, it was somewhere in between being an applied and a basic research group.

Millman:

Well, essentially, yes, if you have a good applied physics group, you don’t mind if some people stick a little closer to the use, and some people say, “Look, we’re going to need to know a lot more about the physics involved.” So whether it looks like basic research or applied research depends on how you look at it. If you can’t distinguish between the two, that’s all the better.

Hoddeson:

Were they all experimentalists in that group?

Millman:

Dillon, experimentalist, Geschwind, experimentalist. Gilleo was looking at new materials, which several years later led to the garnets, certainly an experimentalist. Hogan was an experimentalist with a very strong grasp of theory, but he’s an idea man, you know, patents and experiments, I think you would have to call him an experimentalist. Nicalaudis didn’t have a Ph.D. He went back to Greece a few years later. He was an experimentalist. Prince and Rowen were experimentalists, and Sauer was almost like a technical aide. Yes, all experimentalists. There wasn’t a theorist in the group.

Hoddeson:

OK, let’s skip to this contact physics group and Germer, what are they doing?

Millman:

The contact physics group you could also call an applied physics group because they were trying to understand the physics involved in the operation of the telephone relay. Nowadays there is a lot more electronic switching, but we still use lots of mechanical relays to close contacts. Open and close contacts. Still big business. Anyway, the relay opens and closes, and after some thousandths of openings and closings, it fails in some way. Due to some deposits on it, you may not get a closed electrical circuit when the relay closes. It corrodes or whatever it does. And Germer was trying to understand the physics of what happens, the transfer of material from one electrode to the other, and what would you do to avoid it, and so on. So he was working on the physics, trying to understand the mechanism of the short arc, short arc being low voltage, very, very low voltage, and two electrodes very close together as they open and close. What goes on when it opens and closes? How is it affected by circuits, by inductances you add? Timing of the amount of transfer of material and so on, what’s what Germer was doing.

Hoddeson:

Was he still active in the lab?

Millman:

Oh yes, quite. As a matter of fact, one of my first duties was to strengthen his very small group. When I came in Kisliuk wasn’t there yet. There was Haworth with a physics education only at the bachelor’s level and no knowledge of modern physics. Reitter was essentially a technician, good at mechanical things, building apparatus. J. L. Smith was excellent with electronics and high speed scopes. Germer had been promised to have a reasonably active group and was practically alone, and couldn’t I help build up the group? So I recruited Kisliuk from Columbia. Later on even a better man joined the group, a fellow by the name of Bill Boyle from Canada. He is now an executive director and is very good. He’s been associated with the charge couple devices, CCD’s, inventing even at the level of executive director. He’s very good. So that was that group. That group had existed before I came, but I was encouraged to try to build it up a little bit and get some good physicists in there, and I got Kisliuk and Boyle. Incidentally, Germer several years later gave up that activity. As the Bell Laboratories got involved more and more in electronic switching, Germer, Kisliuk and Boyle published quite a few good papers on the mechanism of what goes on in the short arc. Well, anyway, but later on, they gave it up, and Germer was encouraged to go into surface physics. He went back to his old love of, you know, electron diffraction, and it became a big business, and you know this field of LEED, Low Energy Electron Diffraction, which is an important tool in studying surfaces. Well, anyway, Germer helped in building that up about three years before he retired.

Hoddeson:

How long had he been away from diffraction?

Millman:

Oh, many many years. When I came to the Physical Research Laboratory he had been working many years at this contact physics. It wasn’t just fresh. And he had worked after ‘52 for many years. Quite a few, yes.

Hoddeson:

Here, under general physics, we have only K. K. Darrow. Was he around?

Millman:

No, he wasn’t around really. He was the secretary of the American Physical Society. He had to report some place, and the Physical Research Laboratory was a logical place. General physics, just a name, didn’t mean much.

Hoddeson:

I see. Did he ever come to the lab?

Millman:

Seldom. He had an office at Columbia. That’s where the Physical Society was, at Columbia. He hated to come to Murray Hill and so on. He had some kind of an office at West Street, at the West Street Laboratory, for a while, but he didn’t do any research work for the Laboratories for many years. At first he was writing good articles, in the BSTJ, yes. But he never did work on specific research topics, either basic or applied research, for Bell Laboratories, just writing, and then became secretary.

Hoddeson:

Well, he played a function in the larger community certainly; I don’t know about in the smaller community. All right, now the physical electronics group under McKay —

Millman:

He’s currently the Executive Vice President. Now it includes people like Hagstrum there. Later Hagstrum became the head of this group. These people had to do with what they call gaseous electronics, and these people had to do with very high vacuum techniques. You know, ultra-high vacuum was essential for clean surfaces. But the most important member of this group is Homer Hagstrum.

Hoddeson:

I notice that you certainly have theoretical people mixed in with experimentalists; Conyers Herring is in that group.

Millman:

Yes, before 1956 there was no theoretical physics department. It was believed that the right thing is to scatter the theorists, to interact with experimentalists. It’s still a good idea. But by 1956, we had quite a few theorists, and the theorists essentially promoted the idea that, although there’s nothing wrong about the theorists interacting with the experimentalists and this should go on, but nevertheless it would be nice to have a theoretical group, who would have the charter, the responsibility of thinking in terms of what’s good for theoretical physics. I mean, you know, whatever it takes — if you want to take in post-docs, you want to have seminars. In other words, there is a group concentrated on theoretical physics. It wasn’t intended that all of the theorists in Bell Laboratories join this group, but those who were eminent enough and wanted to could join. So they had this common interest. They began having, oh, daily teas, you know in the tea room, and essentially revolved around the subject: what Bell Laboratories ought to do from the point of view of theoretical physics — invite summer people, or post-docs in physics and so on.

Hoddeson:

Whose idea was that?

Millman:

Essentially Herring’s. Herring and Anderson were the chief proponents of this, and they were the first department heads. The idea was, the department head was going to be a rotating chairman. I mean, he wasn’t going to be there for a very long time but only for a period of about two years. So Herring was the first, and Anderson was their next, and then Lax and others.

Hoddeson:

Was this the period when lots of summer visitors started coming?

Millman:

Yes, yes. Some were there before too. I think ever since ‘52, ‘53 there were people like Kohn and Luttinger. They used to come almost every summer, you know, it was a tradition for them to come in the summer for a month only. They managed to arrange so that they overlapped, because they always worked together for several years, and they arranged so that during the month of July they were both there, or overlapped at least; maybe one of them stayed longer, one went elsewhere. Yes.

Hoddeson:

You would say this increased starting around ‘52?

Millman:

Yes. I don’t know whether you could say somehow or other, ‘52. I don’t know whether –- in other words, one should not attach too much importance to the fact that in 1952 I became Director of Physical Research and things happened. I inherited quite a tradition of basic research, theoretical physics, and so on.

Hoddeson:

Would you say by this time –- see, the questions I’m asking you are sort of the tail end of my story that I’m trying to write, and of course it’s always helpful to see the end of the story when you’re working on the beginning. And the question I want to ask you is, by this time, would you say that a solid state group had been established in a solid way?

Millman:

Oh, I think even before.

Hoddeson:

Well, they started it in ‘45, it took a few years.

Millman:

Yes, they started –- there was some activity before the war.

Hoddeson:

It was not called solid state.

Millman:

Well, that’s right, but there was some activity there. Shockley was there and others. But it was felt that during the war some solid state devices were used, mainly the silicon detector of radar, but with very little understanding of what made it work. You make a whisker-point contact and pass a current through it and it worked, and so on, it was just cook book stuff. And the feeling was, I gather –- I wasn’t involved in that case, I didn’t even join the physics group –- but as I gathered, the feeling was, for one thing, it looks like pay dirt in there. Something works and possibly –- Shockley had in mind, maybe it might be an amplifier or something or other. But here is solid state physics with all these possibilities, and we know so very little about it So the group was set up under Shockley, later divided up into two groups under Shockley and Morgan. This was the solid state physics group. So I think the solid state physics group dates right back to 1945.

Hoddeson:

Yes, it does, that’s exactly when it dates from.

Millman:

That’s right. And incidentally, you’ll find that Brattain speaks with great nostalgia about the functioning of that group, particularly prior to the invention of the transistor. Great admiration for Shockley as a leader. Now, there was some falling out afterwards, because of the invention, with the publicity, and the chief problem was that actually the transistor was invented only by two, Bardeen and Brattain. But there’s very little doubt that Shockley played a very important role. He was the group leader, and Brattain used to speak very admiringly about the way Shockley ran weekly get-together seminars. He’d get the group together, the people involved in the materials and people who were involved in the physics, and put down on the blackboard, “Now, what did we learn? What do we know? What else needs urgent doing?” And it was obvious, nobody had to say, “You do this, or you do that,” it was sort of obvious. “Well, OK, I can do that and you can do that.” He speaks with great enthusiasm and nostalgia. He had his office across from mine for many years, and I got a good feeling about the group spirit that was there.

Hoddeson:

We’re resuming now, after a short off-the-tape break.

Millman:

About lasers, you know, it was obviously Schawlow and Townes who came forth with the “bible” on what it would take to make the laser work, and Javan is the one who invented on paper, in PHYSICAL REVIEW LETTERS the year earlier, 1959, what the neon laser ought to look like. There were experiments to find out precisely what conditions are needed to make the neon laser work, because he thought it would be very critical, so he did a lot of experimenting — he and Bennett. But the man who observed laser action for the first time is neither one of those three principals. It’s Maiman. Quite an ordinary kind of physicist, by comparison. But nevertheless you have to give him credit, he was the one, you know, who first observed laser action. So there is an analogy there, you see, between the laser and the transistor. There’s no question that the original invention of the point contact transistor was by Brattain and Bardeen. But Shockley played a very important role in leading the group, you know — motivated, and later on of course was the inventor of the junction transistor, and the junction transistor is much easier to understand theoretically. It’s a nice, good, one-dimensional problem. Well, anyway, go ahead.

Hoddeson:

I don’t have such a good picture of what the main thrust of the solid state group was in this period.

Millman:

Well, I think as you’ve recorded, initially those two groups, physics of solids and transistor physics, were one group. It’s all solid state physics. First of all, they grew, because of the importance of the transistor and magnetism, so it certainly became large. And it split up, partly because the transistor and semiconductor research was so big by itself, that in its own right it deserved a sort of a group paying attention only to that part of it. But in addition, there was this other motivation of overcoming friction between Bardeen and Shockley, and so they got the Morgan group organized, which included a lot of high level solid state physics, not semiconductors, although you have Brattain, you see, for personality reasons, you have Brattain in Morgan’s group, although Brattain continued to work on transistor-type activities, particularly surfaces. He was very much interested in the surfaces of semiconductors and electrolytes and so on. Brattain is the exception here, but the other people, as you know, are not primarily semiconductor researchers. Anderson is a theorist, you know him. Benedict was recently imported and so on, and he was in solid state physics. Bond, who has retired from Bell, is now at Stanford. He is a very good crystallographer, very good in techniques and so on. Bozorth of course achieved eminence in magnetism, and there were people like Nesbitt and Williams with him, although Williams doesn’t show reporting to him. H. J. Williams did a lot of work on magnetic domains. As a matter of fact, he had some collaboration with Shockley in the earlier stages. But Williams was involved in this. John Galt was relatively new, and he worked on things like losses in ferrites, on magnetic domains, domain motion, understanding the physics of a domain moving and what losses are associated with it, and later on, he changed from this field into cyclotron resonance of semi-metals — semi-metals like bismuth and graphite. Well, there was Geller, a crystallographer brought in. He has since left. He’s now at the University of Colorado. Anyway, Geller, Holden — Holden is retired, of course, but Holden, yes, at that time he was working on spin resonance of — well, he was in crystal growing and in ferroelectrics and also some complicated organic material with an unusually narrow line of spin resonance with a g of 2, and used for calibrating magnets.

Hoddeson:

Did he get a degree?

Millman:

No, he doesn’t have a Ph.D., no. He’s got a bachelor’s from Harvard.

Hoddeson:

I heard he was in the accounting department when he first came to Bell.

Millman:

For a short while, yes. Interesting history. Hal Lewis, you know, he’s out at Santa Barbara. He’s a theorist. He was interested in superconductivity. I remember having heard his lectures. Matthias of course you know. Merz has after several years left for RCA. He’s running the Swiss Laboratories. He was around for half a dozen years. Merz was interested at that time in ferroelectrics, particularly barium titanate and so on. Merz and Remeika as a matter of fact. Remeika was the fellow who would grow the crystals, and Merz was trying to study the physics of it. Murphy had some past history — he was interested in dielectric properties of ice, something like that. Schawlow of course was brought in from Columbia, and he was at that time working on superconductivity. And Williams we mentioned; he was working on magnetic domains. Mrs. Wood, crystallographer; Yager, spin waves, and so on, so there was a lot of solid state activity, good physics, but not semiconductors except for Brattain.

Hoddeson:

And then the other group was really focused on semiconductors, is that right?

Millman:

Well, I think so, yes. Walter Brown was working on super-conductors; Bob Fletcher and Goucher were involved in things like quanta, in light, I mean, interaction of light with semiconductors; Haynes also, yes, both of them passed away since. Hornbeck was involved in the famous Hornbeck and Haynes experiment about, oh, about holes and electron recombination and so on. He is now Vice President at the Bell Labs. Moore, an electronics expert. Machlup was a post-doc, I guess, call him temporary, a post-doc employed as a theorist. Montgomery, also a circuit man. Gerald Pearson is a vary inventive kind of a fellow, an experimentalist. He was the one who later with Fuller invented the photoelectric solar call. But he was involved in thermistors and devices of all kinds, but he’s not — for some reason or other — did not come in on the inventing of the transistor. Could have been, but he was involved in other activity at that time. Prince was involved in semiconductors, later moved into the development department. Thornton Read, you may see sometimes. Read is the one who, at that time, was working very hard on dislocations. He had some book on dislocations. There’s some Read dislocation process. Dislocation, screw dislocation. Read was also the inventor of the Read diode. Have you heard of the Read diode? You know, it’s a very important thing, the outgrowth of the Read diode is what they call the impatt tubes, solid state amplifiers that are going to be used in this wave guide communication at very high frequency, in the region of, say, 40 to 100 gigaherz. This is the one where the propagation is in the circular electric mode. You have a two-inch pipe, but one mode going. Actually, the pipe can transmit all kinds of modes, it’s very big, but properly launched and, if the pipe is perfectly cylindrical and doesn’t have sharp bends and has inserted special mode killers, you can maintain that mode and propagate it. It is a mode where the losses decrease instead of increase as the frequency increases. George Dacey was involved in semiconductors, yes, he is now Vice President. Ralph Logan is a Columbia graduate who is now in physical research in the solid state electronics laboratory, 115. Jon Ross is another man; Dacey and Ross invented field effect transistors. Ross is also a Vice President. Van Roosbroeck is a theorist; you’ve met him?

Hoddeson:

I haven’t met him. I’d like to ask you a general question, before going on to the next area, which is going to be the CDT. Just a general question about having the department heads in every case be people who are actually working in research.

Millman:

Doing research, right.

Hoddeson:

Would you say that’s one of the strengths?

Millman:

Yes, I think it’s one of the strengths. I think it makes that kind of position an ideal job for people who have a little interest in administration. They have some influence and have prestige on the outside. At the same time, they do not give up research. You have an outstanding example in Walter Brown. Have you heard of him?

Hoddeson:

I’ve heard of him.

Millman:

Well, he has the Bell Labs responsibility for the Bell-Rutgers Tandem Accelerator Laboratory. He is the idea fellow who has been several times offered positions of higher administrative level as a director. He’s an excellent administrator, too, but doesn’t want it because he enjoys his position of having his hand in, in his own research, and collaborating with scientist, on a scientist’s level, and not just being involved purely with administration. I think that’s something which is encouraged, and if you find a department head who’s not himself doing research, he is very likely in the development area where the department heads have so many responsibilities, and are involved in frequent conferences and with schedules affecting Western Electric production and so on, that generally the department head does not get much chance to do his own technical work. But in research this is the norm, not the only norm, I mean, I would say it’s at least true of 95 per cent. Now, you might find an occasional older scientist who is not doing much on his own, and is a department head. This is an exception. I think that’s a very good arrangement.

Hoddeson:

I haven’t checked out how far back this tradition goes, but it must be quite a long ways.

Millman:

Well, you’ll find that Stan Morgan, at this stage of the game, was not doing research. He is a very good chemist. I think when he brought in to be in charge, or co-department head, he stopped doing research -– it’s quite unusual. At one time before the department broke up, it was the Shockley-Morgan department; they were co-department heads. And when I was appointed department head in ‘47, Pierce and I were co-department heads, so we followed the precedent of Morgan and Shockley.

Hoddeson:

There, my impression is, the responsibilities were divided into the intellectual and the more or less administrative.

Millman:

Well, that’s probably true, in the Shockley-Morgan case. It’s certainly true that obviously Shockley was much the superior scientific leader, but they also had that division of chemistry and physics, because it was widely thought that materials played a very important part, you know, in this whole research. You now only had to have pure materials, you’ve got to knowhow to put in the right amounts. Materials are very important things; to have somebody with a chemical feeling, what can be done, what it takes organization-wise, was also important. So I would guess that’s why Morgan was put in. But ever since, I know, that is, from the time I came to the Bell Laboratories, when they were co-department heads of that group, Morgan did not do research on his own. But prior to that he was a chemist. Shockley always did his own. Oh, yes, Shockley, you couldn’t keep him away. Working, producing papers and so on –- even at the director’s level. Well, Shockley I would say is not typical of a director. I mean, the typical situation at Bell Laboratories, if you get to the director’s level, you stop doing research of your own. You may coast for a year or two, you know, finishing things up, momentum of previous work, produce another paper or so, but there are some exceptions. The only exception that comes to my mind right now are people like Patel who’s a director and quite active himself. But the norm is not to do research. The department level is the right level to have some administrative influence and prestige outside, inside. I think, from my point of view, it’s the ideal position at Bell Laboratories.

Hoddeson:

Now I’m going to ask you about educational activities at the Bell Laboratories, both learning amongst the people in the research group, and about their function in teaching other people.

Millman:

With emphasis on physics, you mean, or in general?

Hoddeson:

Well, with emphasis on physics. I know that you had quite a bit to do with CDT. You began telling me a little bit about that last time.

Millman:

Let’s see, the CDT got started, I think, in 1948. It was nicknamed Kelly College. This was not intended for researchers who were coming in a research area, but more for engineering graduates coming to the development area. The objective was to give them some, maybe additional training, particularly in the knowledge of communications, Bell System technology courses, maybe also a little more advanced work in analysis or mathematics — whatever it was thought they would need in order to have a more creative career in communications. In 1952, at a time when I became Director of Physical Research, there was a strong feeling on the part of Fisk and Kelly that solid state physics was going to play an increasing role in the communications, not only for people who will design devices such as transistors, diodes and so on, but people in the systems who would use them, so the idea of some acquaintance with solid state physics would be highly desirable. But the actual fact was that engineers graduating even from the very best schools at that time, had no solid state physics education and very little atomic physics. So, in 1952, there was a move to up-date the CDT training that the entering engineers were getting, and I particularly was asked to do two things: first of all, to form a little committee to help decide what kind of physics education these engineers ought to get, and then, for the next couple of years or so, to actually be the angel to staff these physics courses that we proposed, to find lecturers and instructors to teach these courses. Well, the first job was over within a very short time. We formed a committee, and I’ve forgotten who was on that. I know Conyers Herring was on the committee, but maybe several others. We decided that they would give the entering engineers three courses in physics. One was atomic physics, another one solid state physics, another one physics of waves. In each case, the classes were on the order of 100 or more. We ran it on the basis of lectures and then small sections, so we needed lecturers and instructors. And I had, for a couple of years or so, had the responsibility of finding instructors. These were mostly Bell Laboratories people who would like to teach, and there were quite a few people who wanted to teach, Schawlow one of them. He wanted to give a course in solid state physics. And people like Holder and Hagstrum, and others. There was no great difficulty in finding them.

Hoddeson:

Did they do it on company time?

Millman:

Oh yes, sure, this was part of their job. And there were, I remember, two conditions to meet in order for me to be interested in them, aside from having been recommended or thought of that these would be highly qualified people. They had to want to teach, volunteer to take the time to teach. Their arm wasn’t being twisted, “Wouldn’t you like to do this for the good of the company?” Of course we told them this is a very important contribution, but nevertheless, they had to want to teach. And the other requirement, even if they were teaching, they had to get the approval of their supervision, so they had the backing of their supervision because the supervision was responsible for the merit review, salary review, and we didn’t want them to teach and at the same time their department head saying, “Well, the damn fool wants to teach, that’s all right, I won’t be in his way,” grudgingly, “Let him.” No, they had to see the values and be enthusiastic. There were three different courses, and there were different lecturers for the three different courses, and in each of the courses there was one lecturer and maybe four or five instructors, and the turnover was pretty big because, on the average, any given physicist was involved in teaching about one and a half times; some thought once was enough, others repeated the experience.

Hoddeson:

Was it a one semester experience?

Millman:

Well, one trimester I believe, or something like that, or the equivalent. I think it’s one semester, yes. It was all done in one year. In one year the students had to take these three courses. But nevertheless, I would say that even though there were quite a lot of instructors to sign up, it was not really difficult. That essentially expresses the reservoir of people in physics wanting to teach. A little, not a full teaching program, you know. But this didn’t last that long, because at least for other areas of Bell Laboratories it was a little more burdensome to find instructors, and so later on NYU was brought in to help with the teaching, and they even established an NYU campus in Murray Hill, and a whole department was established at Bell to administer the Communication Development Training program at Murray Hill. John Shive was for many years the director of communications training, and so on. So we essentially got out of this, and of course later on we phased out of NYU too, and thought it was better for the people who are going to get training on the level of master’s degree or higher to diversify and have them get the training in various universities, and not limit themselves to NYU. Well, that’s a separate story. Does that help?

Hoddeson:

Yes. Now, what about other educational activities? In your department there were journal clubs, or one journal club, at least.

Millman:

Yes. In general, I recall, we used to run — aside from local seminars that did exist and could exist in various groups, such as theoretical physics seminars or magnetism — there were, generally on Fridays, three principal seminars or colloquia. Once in three weeks there was a solid state seminar. Once in three weeks there was a journal club, and as a matter of fact, it still goes on. And finally the third week, there was going to be a more generalized talk in the auditorium, where people would give a talk which would not be directed to the specialist but more to the general audience. By the general audience we were hoping to attract people from all over, scientists and engineers from Bell Laboratories as a whole, to come and hear of new research, a talk not aimed at the specialist but not a popular talk either. On a technical level, but not for the specialists. It turned out that these general talks, in time, kind of faded. It got to be more and more difficult to get these formally prepared talks and so on. But the solid state seminars and the journal club still go on just as strong as ever. The general talks finally got revived a few years ago, in the form of generalized colloquia with invited speakers and so on. But it’s not held regularly every Friday or so. And it’s not limited to research — well, maybe it’s research, but not limited to physical research. Now, are you interested also in continuing education that goes on at Bell Laboratories?

Hoddeson:

Yes, I’d like to know about that. Actually, there’s one question. In the sixties Quin Luttinger gave lectures on solid state physics as a consultant. Was that typical?

Millman:

Well, I think he was involved in — yes, he was giving some course, I guess, in topics in solid state physics or something. What I was referring to is a little more related to the current very massive educational activities that Bell Laboratories has been involved in the past half dozen years or more, continuing education, which applies to all members, all professionals at Bell Laboratories, and doesn’t single out the development areas, and includes Ph.D.’s and people with masters, essentially encouraging people to take at least one course a year, or two courses a year, so as to combat obsolescence. It turns out that the research people, particularly the Ph.D.’s in research, are not great participants, and that’s natural because they feel that in the process of doing research they keep in touch with the literature, so that they’re not as heavily involved. But it’s all done on a voluntary basis. Research does furnish instructors for those courses, but not proportionately, not much more than other areas, maybe slightly more. It’s not that heavily involved in teaching, but when Luttinger gave his course, it was more for the research people rather than for people in the development area. For the development people it was maybe too specialized. But some of our people have given courses, like in band structure of solids and so on.

Hoddeson:

By research people at Bell Labs, for the others.

Millman:

For the others, yes, and non-research people could take them too. Participation in course work, in continuing education, as I recall, laboratory-wide on the whole, is roughly on the order of, oh, 50 per cent.

Hoddeson:

Are any records kept of who attends?

Millman:

Oh my goodness. Well, you’d have to go into the continuing education department, if you’re interested in that. But records — what is kept is when people have completed a course, and there are no records generally kept of failures. There are exams but no grades, pass or not, but no grades.

Hoddeson:

Did Quin give an exam in his course? Hillman: I don’t remember. He was not formally involved in these continuing education courses. In Quin’s case people like Geschwind and others felt it would be great to have Quin give a course in advanced topics, and so he did it for a couple of years.

Hoddeson:

OK, let us now fill in some of your background that we left out of our discussion so far.

Millman:

Go ahead, what would you like to fill in?

Hoddeson:

OK, now, you were born in Poland in 1908?

Millman:

That’s right.

Hoddeson:

When did you come to this country?

Millman:

’22.

Hoddeson:

I see. And where did you settle, New York?

Millman:

Yes, in Brooklyn, yes, New York City, Brooklyn, with relatives. I lived for many years with my brother until I got married.

Hoddeson:

And you went to New York City schools?

Millman:

I went to — well, New York City high school. I didn’t spend that much time in grammar school; it took less than a year until I got into high school. I spent most of my time in night school. Night high school. One semester in day school, then moved to night school, and went to work during the day.

Hoddeson:

What did you do?

Millman:

Oh, after some odd jobs here and there, I, most of the time, worked in my brother’s ironworks shop; I was an ironworker while I was going to school. And then spent about half of my college education in City College at night. At that time, City College had a branch in Brooklyn and I went to the Brooklyn branch of City College at night, and then just at the time Brooklyn College became established as a separate school, I moved to City College uptown in order to be a physics major and take elective courses. You couldn’t get much at the Brooklyn branch, so I transferred uptown and spent my last two years, during the day, at City College, and after than went to Columbia for graduate work.

Hoddeson:

What was your focus at that time in your physics courses, or was there any?

Millman:

See, I went to City College with the thought that I might be a math major. My interest was to teach math in high school. But as I kept on taking — and I knew I was going to like physics, didn’t take physics in the high school because they had a terrible teacher, so I deferred it to college. But as I got to take some elective courses in physics, like mechanics and so on, I got to appreciate the value of mathematics as a tool for solving physics problems, more, even more than as a discipline in itself. So by the time I got through with college education, I could have been called either a physics major or a math major, either one, but I actually converted into being called a physics major, in other words, took more courses in physics, so I graduated as a physics major. Then I went to Columbia to do graduate work in physics, mainly because I wanted to teach, and the high school system required one year of graduate work. You didn’t quite need a master’s degree. So I went to do first year graduate work and picked up a degree; all you needed was the extra $20. The requirement was to pass some exam, but it was an easy examination. But there were no jobs. I was married at that time. My wife was working. I didn’t know whether I would be good enough to get a Ph.D., but there were no jobs, so I kept on studying.

Hoddeson:

Whom did you work with?

Millman:

Rabi. But I did get a degree in a reasonably short time, four years from the bachelor’s, which was pretty good for Columbia.

Hoddeson:

What did you work on?

Millman:

I worked on nuclear moments, molecular beams, hyperfine structure.

Hoddeson:

That must have been very exciting at that time.

Millman:

Yes. I was on the ground floor when Rabi invented or thought of this molecular beam magnetic resonance method for direct measurement of nuclear magnetic moments.

Hoddeson:

Was there much discussion about it going on at the time?

Millman:

I remember, my apparatus was modified from atomic beams to nuclear beams within two months. Kusch and I got the first resonance, lithium resonance. That was a very exciting time. The actual discovery or invention of this process, Kusch and I didn’t have any part of it, actually came about when C. J. Gorter from Holland came to this country and visited Columbia, particularly Rabi, and he was essentially bemoaning the fact that he had tried to get resonance in solids. This is spin resonances and so on. And failed because the techniques were not that far advanced, also he might have used some material with the wrong relaxation time –- at any rate, it wasn’t successful, and Gorter came to discuss this with Rabi. And, I gather, out of this came the thought, why don’t we use these resonance ideas in molecular beams? So Gorter had some hand –- I think Rabi in his first publication acknowledges discussions with Gorter on this, thanking him for the discussion -– maybe Gorter felt that he should have been on the paper or something. I gather there was some feeling of that kind, but never mind. I wasn’t involved in that process. I mean, I wasn’t even in the office. But we heard about it soon thereafter, and Rabi became all excited about getting the molecular beam apparatus which we had geared up for atomic moments. Now, nuclear moments are much weaker than atomic moments, therefore you needed much stronger deflecting fields. So we had to build magnets going inside the vacuums with very strong gradients to deflect the beam in one direction, then back, refocus, and then, if you now put in between a magnetic field of such magnitude that the frequency of precession of the nuclear moment is the same as the radiofrequency field in that region, we have a resonance condition and spin-orientation change, and the beam doesn’t come back to its normal position, so you get a drop in intensity –- and of course this was the way you knew that you had magnetic resonance. So I was involved. Incidentally, I have just sent in some biographical material to Lehigh –- I’m apparently about to get an honorary degree from Lehigh on October 13th, so they asked me to send biographical material. I’ll give it to you.

Hoddeson:

Sure, that will be very helpful. I’ll put that in your file. Your AMERICAN MEN OF SCIENCE biographical sketch indicates you were an assistant professor before you got your Ph.D. at Columbia.

Millman:

No, I wasn’t an assistant professor, I was an assistant. In fact, I was not even a regular teaching assistant at Columbia. I did do teaching in Columbia, I did it in their, what they call general studies or evening. Von Nardroff is the one who gave lectures and he had assistants. So I was not a regular teaching assistant at Columbia. In other words, I was a paying student, had no fellowships and no teaching. But the last two years of my graduate work, I did get paid by the hour, depending on how many hours, I think it was $2 an hour for teaching. I think it was eight hours a week or something like that, teaching at night. So I did teach at Columbia, those years from ‘33 to ‘35. I also did a couple of summers at Brooklyn College, ‘33, ‘35, some evenings at Brooklyn College. I did do some part time. It wasn’t till 1939 that I got a full-time teaching position at City College, and later on transferred to Queens College. No, I never was professor. At the end of the war, 1945, I had the problem whether I should go back. I had essentially had a tenured position at Queens College — to go back to Queens College, or accept offers like Bell Labs, General Electric Co., and other places.

Hoddeson:

Weren’t you at the Radiation Lab for a while?

Millman:

Oh yes, during the war, yes. That was ‘42 to ‘45. Yes, I was yanked out the middle of ‘42, March ‘42, because Rabi decided that it would be a good idea to have a kind of research laboratory on radar, particularly for short wavelengths, 1 centimeter instead of 3 cm or 10 cm at Columbia, away from the main center at MIT. Kellogg, Kusch and I got the Columbia Radiation Laboratory started in 1942. And I was there until the end of 1945. I was associated with something which, if you were in the business of tubes and electronics, you might appreciate. You know what a magnetron is?

Hoddeson:

Yes.

Millman:

A magnetron is the heart of radar, sends out pulses, but I was the inventor of a special kind of magnetron called the Rising Sun Magnetron, which is particularly good for very short wavelengths. Doesn’t have straps in the anode structure, and it has lower loss. That was my baby. I might say, it’s in the course of our work at Columbia on these magnetrons which got us in close touch with Bell Laboratories, particularly with the Fisk group, Jim Fisk had a group working on 3 cm in magnetrons. The magnetron that I invented, the Rising Sun Magnetron, was going to be developed by Bell Laboratories for manufacture at Western Electric Co. Bell Laboratories did indeed take it over for development. It never actually got into full production because of the end of the war. It was used enough in experimental bombers to look at the terrain in New York City and find — indeed, you get extra resolution by going to 1 centimeter instead of 3 centimeters. So it was tested out experimentally. It was never used in war operation. It was intended to help fight the Japanese war, if we were going to invade Japan, so it was never used. But we were very close to the Bell Laboratories people working on magnetrons and tubes, and it was this association that led Jim Fisk to ask me, after that, whether I would want to join the Bell Laboratories. In other words, he recruited me.

Hoddeson:

And now we’ve come full circle. This is where we began last time. I’d just like to ask one more question to help us sum up. I keep coming across general statements by Slater and others on the important role of the war-time microwave radar work to the post-war development of solid state physics.

Millman:

Well, the microwave spectroscopy, initiated mostly by Townes while he was still at Bell Laboratories; later on he moved to Columbia –- the microwave spectroscopy techniques, detectors and transmitters and Klystrons and so on was –- it looked like, kind of, a development effort for a purpose, for radar, for fighting the war; engineering, and so on. But the boost that it gave to pure science, by virtue of having those kinds of tools –- I mentioned earlier, for example, Gorter was looking for resonance and couldn’t find it. But the advancement of these kinds of techniques, you know, was a great boon. One of the things that came out -– well, I don’t know whether NMR, nuclear magnetic resonance, of course, was invented through it. Rabi observed it for the first time; but that was a specialized situation, molecular beams, but to get the nuclear magnetic resonance in a little bit of material such as semiconductors or other solids or water, NMR as it’s known, was done after the war, by Purcell and Pound, and by Bloch and Packard. Well, the boost that that kind of technique gives you to understanding solid state physics is enormous. So, though in some sense it looked as if during the war science stopped, in a broader sense, because of the increased attention to instrumentation and tools and microwaves and so on, it probably promoted the cause of science, such as microwave spectroscopy, masers and then lasers, and so on.

Hoddeson:

Also, the focus on the crystals, the rectifiers, seemed to do two things: techniques for producing good crystals –- relatively perfect and pure crystals –- were developed during the war too?

Millman:

Yes, oh yes, well, that –- but more so, even after the war with the invention of the transistor. As I’ve said over again, one of the things that the transistor did was, not only its effect on technology, you know, and on world economy and so on, and brought about computers and the possibility of going to the moon. What would you do without the transistor? But it also affected the science of solid state physics. The fact that you had at Bell Laboratories a place where it was important to get pure materials -– you had Pfann’s zone refining and so on. Here, you see how much more you can understand about various electronic mechanisms –- transport mechanisms or whatever it is, traps and so on –- in solid state physics if you have control of the materials. You know how to make it pure. You know how to inject impurities of the right kind. So the boon to pure science was tremendous, even if you’re not interested in anything else but knowing more about solid state physics, there was a great boost by virtue of the invention of the transistor. And when Slater, I think, talks about microwaves, he has this kind of thing in mind. I mean, the subsequent advances, the subsequent much more rapid advances that come in physics because of the presence of these important tools.

Hoddeson:

Yes. Yes.

Millman:

OK, we did it.