Oral History Transcript — Dr. Amnon Yariv
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.
See the catalog record for this interview and search for other interviews in our collection
Amnon Yariv; January 28, 1985
ABSTRACT: Doctoral thesis on 2-level solid state masers at Berkeley in 1956-1959 through the beginnings of his work on integrated optics at California Institute of Technology in the mid-1960s. Research at Bell Telephone Laboratories on noise, parametric amplifiers, doped crystal lasers, and semiconductor lasers, and his mode-locking studies at Watkins-Johnson and Lockheed. Role of consultancy at Hughes Aircraft Co. on research in phase-conjugated optics.
Yariv:Shall we talk about Berkeley?
Bromberg:Yes, let’s start talking about your electrical engineering studies at Berkeley.
I had received a degree in electrical engineering at Berkeley in 1954. It was a conventional electrical engineering degree with an emphasis on microwaves, influenced probably mostly by John Whinnery, who was one of the better known professors at Berkeley. And John Whinnery was working on traveling-wave tubes. Those are the precursors of today’s free electron lasers, very much the same kind of device but a different name. I started in 1954 doing my graduate work for a Master’s degree in Berkeley and started working on a thesis in traveling-wave tubes, and during that period, I built an experimental tube and investigated it and really decided I was not interested in the topic. This is a topic—traveling-wave tubes—that people like Kompfner that I see mentioned further down and John Pierce at Bell Labs were responsible for. Traveling—wave tubes you know are used today as the wide-band amplifiers in communications satellites. They were quite important. So I started working on the topic and decided after a year or so that I really wasn’t interested in continuing working in traveling-wave tubes for my Ph.D. I found the range of concepts involved limited. I guess it didn’t touch me, it didn’t resonate.
So I was shopping around, that’s about how ‘55 or so, after a year, year and a half, having written a thesis, for a topic suitable for a Ph.D. topic. And one thing about the education though, various concepts, which later on I used very heavily in my recent work-concepts like coupled mode approach to solving problems (in integrated optics)—I picked up during that period. John Pierce once introduced them to the study of traveling-wave tubes. Of course the whole idea of phase synchronism—what we call today phase matching—is the Key idea in traveling-wave tubes. So the background was probably invaluable, in some ways, later on. But to go back, about ‘56, ‘55. I was shopping for a thesis topic and happened to go to what was called in those days the “tube conference”. It was a small conference, you might have heard about it, held usually in June or July in different places. It was a “by invitation” conference, very small, maybe 100, 150 people, which was started by the people who were then the traveling-wave tube people: John Pierce, Kompfner, Whinnery, Robert Adler of Zenith—a bunch of very creative and original people.
This conference, although called tube conference, was very novel and imaginative conference. Basically, anything to do with microwave communication was acceptable. And because of the influence of the people who were the leaders of the meeting, people like Kompfner and Pierce, they usually brought the best work from their own lab, even when it did not directly involve traveling-wave tubes. So it so happened that in 1956, at the meeting in Boulder, I heard a talk given by George Feher, Harold Seidel, and Derek Scovill from Bell Labs on the first microwave maser experiments.
Bromberg:They had just gotten the first two-level solid state to work.
Yariv:This was the very first announcement of this work at the public meeting. And a lot of important things were first described at tube conference meetings. Jumping ahead, I can name for instance, (that was one of them, the two-level maser), the first parametric amplifier (the low noise parametric amplifier).
Bromberg:Well, Weber announced his version of the maser in ‘52.
Yariv:That’s right. In a ‘64 tube conference in Cornell...Two people, DiDomenico from Bell and I independently gave the first papers on mode locking, ultra short pulses. Today it is called the “device research conference,” incidentally, so it is still alive. Anyhow, to go back, at that meeting I heard a paper on the paramagnetic maser—the two-level maser—and the topic all of a sudden fascinated me, for some reason. I guess maybe possibly the personality of the people involved—George Feher was kind of a legend. He had graduated from Berkeley and went to Bell, and was very very well-known, very imaginative. His style appealed to me. So, I went back to Berkeley and decided that I would like to work in this field.
Bromberg:You don’t remember what the concepts might have been? I mean, were you well-versed in quantum mechanics already?
Yariv:No, not at all as a matter of fact. I had not at that point taken any courses in quantum mechanics.
Bromberg:I see. So the idea of going into a quantum resonance to grab out some kind of a circuit element must have been...
Don’t really know why it touched me. Because I really in some ways lacked the background. But I think probably the people—the style there—and sensing intrinsically that it was a deeper subject than traveling-wave tubes, that there was more to it that it touched on. So, I went back to Berkeley, and told my research advisor, Prof. John Whinnery, that I would like to work in this field. I was in the electrical engineering department. And John agreed, but said, “Look there is a problem. None of us here know anything about the field.” I mean, they knew as little as I did. “And what is more, you know, we have never done this kind of research at EE. We don’t have the equipment, the background; you are on your own more or less.” But he agreed to remain my supervisor, although he could not really in effect support the work the way he did his other students. So I went back—that is ‘56—and for one year, just took courses in physics, starting...Well, I had previously taken the freshman physics that all Berkeley engineering and physics students take, Physics 1, 2, 3 type sequence, but no graduate courses. So I started taking courses in physics for one year. I took Kittel’s solid state physics, I took a course in advanced quantum mechanics from William Nierenberg, who is now the Chancellor of La Jolla, Scripps, and took a course in magnetic resonance form Erwin Hahn, you know the spin echo Hahn, and group theory from Mike Tinkham. So I essentially took all my physics there in one year, a very concentrated dose. And had to resist a whole bunch of attempts from physics professors to convert me, to make me shift over to physics.
I did well in my courses, and some of them were telling me I should switch over to physics, but I did not. And during that year, I started collecting equipment, mostly microwave equipment which consist of: klystron oscillators, wave guides, magnets and Dewars. We didn’t have any of that, so I had to design and build a lot of it. As a matter of fact, we had our own magnet cast, because we could not afford to buy one. So we bought the windings from Varian Associates. You know those microwave two-level masers were crystals in magnetic fields. I also had the benefit of the legendary George Becker the all time master technician who could build and fix anything mechanical and of a talented microwave engineer Fred Clap who helped me build Pound Stabilized microwave oscillators. I more or less decided as a starting point to start building an apparatus which was somewhat like the one Feher had built for the 2-level maser. I looked at F-centers in magnesium oxide which were induced by radiation. So I was taking courses, building a system, and about a year after I started a new graduate student Jim Kemp also joined the effort and we overlapped for a while. He helped me greatly in doing experiments and later on continued experiments with the same set-up.
I wound up doing, for my thesis, a 2-level microwave maser experiment in which you look at properties of spins in magnetic fields. You invert a population by adiabatic fast passage, sweep the magnetic field while a [dc] magnetic field is on and the magnetic field sweeps through resonance. And the net result is, at the end of the sweep if it is done fast enough (not too fast, not too slow), the spin population is inverted. And if the population is then situated in a microwave cavity, there is a pulsation involved after that as the inverted spin give up their stored energy to the microwave field. I analyzed the problem using a formalism which were developed for that kind of physics, but not necessarily for a maser project, by Bloembergen Purcell, and Pound. And observed, it’s in my thesis, some of the intensity pulsations, which are today associated with things like spin-echoes. Much of the work of Feld at MIT, and Abella and Kurnit later on at Columbia and Chicago had to do with the same phenomena in the optical regime.
Bromberg:I think of there not being so very much work on two-level masers, I think of Bell Labs, Westinghouse; now you’re saying you were actually experimentally working on two-level masers, Was there a lot of that going on?
Yariv:No, there was not a lot going on. These may be some of the few, maybe the three experiments. The reason the work died is because by their nature two-level masers were transient type amplifiers: you know they emitted the pulse, and then you had to invert the spins again and start over again. So they were probably limited in their usefulness. The importance was probably more fundamental, in showing some of the basic principles involved, such as inverted population, radiation damping, super florescence and what we call today in lasers relaxation resonance, and as an educational tool.
Bromberg:And then there are two other things that I was wondering about. At this point, is there anything else going on at Berkeley? Hahn and Singer later get into masers and lasers, but they were not doing anything at this point were they?
Yariv:No. Singer was hired in electrical engineering and Singer was hired ostensibly to help me, because John Whinnery felt sorry for me. There I was working alone, nobody understanding what I was doing in the department. There were some people in physics who were in the general area. Those are people like Kittel, Alan Portis, and Walter Knight—actually there was a whole group on paramagnetic resonance working in physics. But they were interested mostly in NMR, in basic studies and not paying much attention, except Alan Portis who was talking to me every now and then. The truth probably at that point, they did not yet appreciate the challenge and the promise of, call it the quantum electron aspect of lasers and masers. They saw that as maybe a more practical direction of magnetic resonance, and were not turned on.
Bromberg:And then also at Stanford they are just beginning around ‘56 to work on masers at the electronics laboratory. Did that have any interaction?
Yariv:Yes, there’s work at Stanford on the ammonia maser.
Bromberg:That was Helmer, I guess.
Bromberg:But then around ‘56…
Yariv:Not to my knowledge.
Bromberg:I can’t remember the name. But one of their really big people starts to build up. By ‘57, people like Siegman, were publishing. They were not in your universe at this point, is really what I’m asking you.
Yariv:No. I am not aware of any experimental work at Stanford on masers and lasers until Helmer. That’s about, I think, that period, is it not?
Bromberg:Helmer on the ammonia beam maser. But Siegman, and these other guys on solid state masers, which I would just have guessed might be more close to what you were interested in.
Yariv:Yes, possibly a little later? I’m not sure of the dates, but you are the historian—maybe at the beginning, theoretical work, yes.
Yariv:I remember at some point some calculation by Siegman’s group of matrix elements in ruby and so on.
Bromberg:Was there much going back and forth? Personal contact is really what I am searching for at this point.
Yariv:No, none. None with the Stanford group.
Yariv:I had some contact with them in my traveling-wave period. That’s because there was a lot of traveling-wave research. Watkins and Johnson, for instance. Watkins especially, who founded Watkins and Johnson and was then a professor at Stanford. But not during the maser period. So that takes us to about, let’s say ‘57, ‘58 when I am actually working and doing the experiments. And my first contact with Bell Labs, besides the initial contact—you the George Feher and so on that got my started—is during this period, where both Rudy Kompfner and Chapin Cutler, if you know both names.
Yariv:Are frequent visitors to Berkeley.
Bromberg:I did not know that.
Yariv:Kompfner just dropping by occasionally. Cutler actually spent a sabbatical year. It could have been ‘57 or ‘58, he spent a whole year there. So I have gotten to know the people, and they became acquainted with my work and more or less encouraged me to come to Bell after I finished. So when I finished in ‘59, December ‘58 is when I left Berkeley, I went directly to Bell, had an offer from them.
Bromberg:Do you have any sense of how interested or not interested Cutler and Kompfner were in the maser at this point?
Yariv:I don’t think they had any great particular interest in the maser, per se. But by ‘58, you see, there was already the beginning of the work on lasers. By late ‘58 Townes and Schawlow had written their paper, and Bell at that point already had a group of people, maybe two groups trying to make the world’s first laser, or as they called it, optical maser.
Bromberg:I would be interested in your own reaction to when you heard about the Schawlow-Townes paper, and what you thought about it or didn’t think about it.
Yariv:I must admit that the Schawlow-Townes paper—I don’t recall the exact date probably was published during the last of my Ph.D. research, and I was so busy working on my experiments, you know, trying to get out, that I knew it existed but I wasn’t struck by it, it wasn’t like a revelation. It was too many other people.
Bromberg:Did Gould’s work get known at this point? I mean, did the fact that TRG was beginning to mount this enormous research program…was that just known in the community?
Yariv:Yes. After I joined Bell Labs, we knew about the work at…what was the name of that company?
Bromberg:Technical Research Group, TRG.
Yariv:TRG, and about Gould, but I was completely unaware of what later on came to be the controversy for credit, and so on. We just knew there was a group that has a government contract to make a laser, TRG. And when I came to Bell, I found two groups already involved in an attempt to make an optical maser. One group involved Bennett, Herriott, and Ali Javan, and they were trying to make a gas laser based on helium neon, which eventually succeeded. Before we leave Berkeley, let me read your questions fast, and see if there is anything here that…OK, so we talked about chief scientific interests. Maybe we covered that but looking back I think that the microwave background and by microwave I mean the concepts of waves and interference and cavities—was very very important. I try to make a point, even in my text book in the introduction, that the two main ingredients to optical electronics are the wave electromagnetic aspect and the atomic physics, and that they are really equally important. And the Berkeley period really gave me in a way, just fortuitously, both backgrounds. I studied waves as a part of the microwave curriculum and then atomic physics and spectroscopy and the paramagnetic resonance as the preparation for the two-level maser work. Just an accident. Then we talked about the paper, talked about Kemp. Kemp was a student who joined about half way through my research, who worked with me for maybe a year. Was very, very helpful, a very bright fellow, and a very good experimentalist. And then he stayed on after I left, and did a thesis—a good one—then went on to Oregon where he is a professor now. And with Hahn, yes I took as I mentioned a course with Erwin Hahn—but otherwise, did not have much interaction, except with his ideas... a term paper that I did for Hahn on spin echoes, probably was my first kind of delving into this whole area, and probably helped chose the topic. So, yes, I knew a little bit about [spin echoes].
Bromberg:The non-linear treatment I guess interests me a little bit because I think of non-linear knowhow being engineering at this point. And I am wondering if that is just a false impression. When I see people talking about non-linear oscillators in the ‘50s and even in the early ‘60s, I think, “well these are the electrical engineers who really learn this kind of thing, whereas the physicists do linear physics.” And so I was a little bit interested to see your explaining these intensity damping by saying you have got to use a non-linear treatment here. Do you think that I am wrong in paying that kind of attention to it? It was this second paper here, on the bottom.
Yariv:“Radiation Damping effects in two—level maser theory”
Bromberg:And then you did another paper later on.
Yariv:Yes. Something about non-linearities... Electrical engineers were sensitive to non-linearities much earlier than physicists because real systems are almost always non-linear. And the engineers have less freedom to choose their problems. A physicist can choose what he wants to work on, an engineer has to solve a problem. If the system is non-linear, he has to account for it. And many circuit elements you know, are non-linear. Also, oscillators essentially are by their very nature a non-linear, and oscillators play such a key role in electrical, engineering. There is just no way to ignore non-linearities. So electrical engineers were forced probably earlier into realizing that the world is non-linear. But they have done it in a somewhat different fashion: they took their circuit theory that deals with resistances, capacitances, inductances, and then said, let’s go one step further and now assume the circuit elements, are not described by constant numbers, say a five ohm resistor but they depend also on the current or the voltage. And that’s the electrical engineering approach to non-linearity. And the kind of non-linearity which is implied in the paper that you are talking about—the two-level paper—was really somewhat different, although at some point the two can be reconciled. It simply says, let’s take the differential equations that govern the fields the atoms and their interaction and not make approximations, and try and solve them more or less exactly. And if you do so, you really have a new way to treat a non-linear problem.
Bromberg:OK. But it is not particularly then an engineering way. It could come just as easily out of the physics year you spent.
Yariv:It does. Yes, that’s right. If you simply insist on solving a certain class of problems in physics without making linearizing assumptions, you have an inherently non-linear problem. And the equations which I solved in my two-level maser, were really the Bloembergen, Purcell and Pound equations. I simply applied them to my case. And the equations are inherently non-linear to begin with, if you don’t approximate them.
Bromberg:Good. I was just really confused, so I am personally glad to get that cleared.
Yariv:So we now are at Bell. The reason I was at Bell was probably mostly due to getting acquainted with Kompfner, and Chapin Cutler, and also of course the early contact with George Feher. George Feher was in the physics department at Bell, and working on paramagnetic resonance. The two-level maser I think was a deviation for him, a detour for him. His main work was in spin resonance in semi-conductors. He invented a technique known today as electron-nuclear double resonance technique: ENDOR. And simply was probably the leading world experimentalist in paramagetic resonance. Had all the apparatus. Seidel was probably the initiator of the project. He brought together Scovil and George Feher.
Bromberg:Seidel, Harold Seidel?
Bromberg:Because I think of him as a kind of an electrical engineer, I didn’t know that he...
Yariv:Yes, very talented and versatile, roaming really over many areas. The kind of person who really probably knew what was happening and the importance of those… You see, he was not a specialist. The other two people were specialists, and he probably, though I’m not sure of it… My guess today that he was the one who said, “Hey, Derek, drop what you are doing and, George Feher, here is an exciting problem that we could possibly do together.” But you will have to check on that.
Bromberg:That’s very interesting. I also never realized he had possibly a kind of organizing role.
Yariv:So I went to Bell and worked for a while on building an Esaki diode, today called also tunnel diodes, microwave amplifiers and oscillators.
Bromberg:Is that because you came into a group? What part of Bell did you go into at this point?
Yariv:I joined the electronic research laboratory, the research department. A group which was at some point headed, at some level maybe, by Rudy Kompfner, and under him were people at that time like Chapin Cutler, again one of the people I mentioned before, and a group which had traditionally done traveling wave research before that. John Pierce was previously also part of the group. The group at this point was beginning to do research in solid state devices, semiconductors, and optics. So I did some work simply looking for something to do, and then joined the group working on lasers—still called optical masers at Bell. At that point, I mentioned already two groups working on masers…oh I mentioned one, that’s the group of Ali Javan, Herriot and Bennett, I guess we are now talking about 1959. And then another group working on a different type, trying to make masers based on some crystals.
Bromberg:Now would that include Schawlow and Devlin at that point?
Yariv:That group includes Schawlow and Devlin and Garrett, Geoff Garrett and Wolfgang Kaiser, and maybe Don Nelson, and Bob Collins—a big group. Trying very much, very hard. And I joined efforts with some newcomers. Our group was in the electronics research lab, they were in the physics department, solid state physics. And we started working on calcium tungstate, crystals with different doping impurities.
Bromberg:Really, at that early stage? Because you know your papers don’t really reflect that. You have a lot going on here on noise and Esaki diodes, but the first papers on solid state lasers start coming out in early ‘62 when you are working with Porto and Nassau and those people.
Yariv:I started working on it in 1960, yes. For the first year I worked probably completely on noise and Esaki diodes. And some magnetic resonance work, then joined, probably in 1960, a new group consisting of Porto, myself, Kurt Nassau was the crystal grower. And we started looking at calcium tungstate doped with various rare earth ions.
Bromberg:OK. Then Maiman’s discovery, how did that intersect with all of this? Did you start with these other people before Maiman, or after, or was that any factor at all?
Yariv:All of this work is going on before Maiman. Do you happen to remember Maiman’s first disclosure?
Bromberg:I think it was July ‘60, when that press conference was held.
Yariv:Oh, July ‘60?...
Bromberg:May ‘6- I think was when he got his first results, but July ‘60 I believe was when he first… The Maiman thing at Bell was a kind of a turning point for intensification of research?
Yariv:Yes, I think so. The work was quite intense before that—I mean, there were probably maybe altogether maybe 10 people at Bell Labs, we mentioned the names, all trying to make the first laser. But Maiman’s work really was a step function in excitement and in effort to do it. So you say, “How did your paper with Kompfner come to be written?”
Bromberg:That was just a bit of curiosity on my part, because for the first time I’ve seen somebody writing a paper with Kompfner in that period, and I just was curious about it. But you know, in all of this I may not have picked out the most important papers, so we ought to keep in mind that there may be other papers we should be talking about.
Yariv:Well we are going down the list, soon we will be coming to a paper “Quantum Fluctuations and Noise in Parametric Processes,” which I think is worth talking about, but that has a bearing on your question about the paper with Kompfner I was interested in noise as you saw, mostly because of my work on tunnel diodes, before that. That’s probably my beginning interest in noise.
Bromberg:Oh, I see.
Yariv:‘Cause the tunnel diodes have turned out to be a low noise amplifier, and so noise was a topic of interest in the beginning. I wasn’t maybe too well aware of why, but back in ‘59, I think it is ‘59, there was the first quantum electronics, the Schawanga Lodge meeting.
Bromberg:Did you go to that?
Yariv:Yes. I was fresh out of Berkeley, and I was invited to go along by people like Rudy Kompfner, and it was not far. I had nothing really to contribute there. So I went to the meeting, and you know, met everybody. You must have heard from other people about the meeting?
Yariv:But one thing about the meeting which attracted, besides almost drowning during the water polo game...
Yariv:We played a volleyball game with Kompfner and a bunch of other people, and there was a Dutchman Bolger, and we played water polo. And I guess the European rule is that if you keep the ball under water your opponent can sit on you until you release it, and I didn’t know it and he sat on me—and he weighed twice as much as I did and finally when I came out I was all blue. Pointing to his bookshelf, one of those volumes is the Schawanga Lodge Conference. And in the Schawanga Lodge there were a lot of papers. Everybody was there you know...
Bromberg:There it is, “Quantum Electronics”, that rust colored volume next to Plasma Physics...
Yariv:And in that meeting...You know who was there?
Bromberg:Yes, the list is in the volume.
Yariv:But there were also papers on parametric amplifiers there.
Yariv:And parametric amplifiers are really the precursors of all of non-linear optics—it’s the same theory, the same topic.
Bromberg:And that was very active at Bell just at the time, Suhl, and Weiss.
Yariv:That’s right, they made a ferrite microwave amplifier. Well, ferrites didn’t survive really. There was another approach, which I think was started, if I’m not mistaken but I could be, at Standord by Wade and Heffner using non-linear diodes, and that’s the approach that eventually turned out to be the practical one... And in Schawanga, Heffner gave a talk, “Parametric amplifiers and their comparison with masers.” P. 269. And here he essentially made the claim that parametric amplifiers were potentially better than masers. The noise performance of masers, or lasers had already been analyzed by Shimoda and Townes, and they had shown that lasers have a quantum mechanical limit to their sensitivity, to their noise performance, due to the quantum mechanical (zero point) fluctuations. Parametric amplifiers were also analyzed. And also their noise behavior was understood, and Hugh Heffner’s paper somewhere here showed that if you take the theory of parametric amplifiers, you conclude that they can have lower noise than lasers. So here…my contribution to the meeting was a small comment that this conclusion that parametric amplifiers are better might, it might be based on the fact that parametric amplifiers were analyzed until then only classically while the laser was analyzed quantum mechanically. Maybe one has to look at parametric amplifiers quantum mechanically before one can make a comparison. I didn’t know what the answer was, but I was disturbed.
Bromberg:And this was a kind of inception of your interest in this particular thing?
Yariv:Yes, just exactly that. So I went back to Bell and essentially started studying what now we would call quantum optics. I tried to formulate a quantum theory for parametric processes. And I did it. And after a lot of struggling, learning a lot of concepts, I essentially came up with the Hamiltonian for nonlinear resonator or more precisely for electromagnetic modes coupled by a strong nonlinearity. This is probably the beginning of the field of non-linear quantum optics, is Hamiltonian a non-learn term—a Hamiltonian is a term which included the products of the electric fields. All I did really was to add the three mode (harmonic oscillators).
Bromberg:Now how did Louisell and Siegman get into this?
Yariv:The way they got into that is the following: Louisell was a very powerful mathematician-physicist at Bell, with a lot of major contributions to traveling wave tubes, a field which I already mentioned. And I actually did the theory at Bell, but my theory, like many, gave the right results—I knew it was correct—but wasn’t very polished. I went to Louisell really for the mathematical rigor. So Louisell re-worked it. When we finished, Louisell went to Stanford and gave a talk on the topic. At Stanford he found out that Tony Siegman had also been working on that problem, independently.
Yariv:Probably was bothered maybe by the same problem. Siegman had used some mechanical model but was on track, and he had some of the elements of our work, so the right thing was to join efforts and write one paper.
Bromberg:That’s extremely interesting. Of course, Stanford was very active in parametric amplifiers—Heffner was from Stanford.
Yariv:Yes. You see, at Schawanga, my question probably bothered them too—you know there was a question, are parametric amplifiers really better?
Bromberg:That would be interesting to know whether they…the inception of Siegman’s work also might have…I must ask him at some point.
Yariv:I really don’t know. But you know, some of those ideas are fuzzy, at the beginning of something…when an idea comes out of the mud and begins to crawl on? It’s not very clear sometimes how and why it starts.
Bromberg:Yes, that’s true.
Yariv:But you know, Siegman, much of his work since then is very much interested in problems of noise and fluctuations. I think in some ways, we are very similar—backgrounds in electrical engineering, heavy emphasis on waves, and then going into physics.
Bromberg:Now I am a little confused. Do the Esaki diodes fit into this in any way, or is this a separate kind of…
Yariv:They don’t fit in terms of the formalism. Oh, what got us started on the whole thing was how did I write a paper with Rudy Kompfner. So, Rudy was at Schawanga—he is probably the one who took me there, encouraged me to go. Rudy and I, at Schawanga, as a result, started talking about noise, noise in general. Rudy, of course you know, was working on traveling wave amplifiers, and one of the most important issues in traveling wave amplifiers were noise, how low a noise figure can one get. Herman Haus, as you probably know, had written a definitive paper those days on noise of traveling wave amplifiers and so on. It was a very interesting topic. My thesis advisor, John Whinnery, had made contributions. So, I was tuned to thinking “noise”, but the formalism here, really the key thing is the inclusion of the term in the Hamiltonian, which is a nonlinear term, a term which is capable, so to speak, of annihilating pump photons and creating signal and idler photons. So it is a term which is third order in the product of the electric fields. The usual Hamiltonian is only second order. So if you only go one more order in the electric field, all of a sudden you have a term now that can create simultaneously a signal and idler photons and generate a pump photon that can create and can do it spontaneously. Because in quantum mechanics you have zero point vibrations. And once I had the term, I had all the results, really. Anyhow, that paper probably is the first paper. There really is nothing before that on nonlinear quantum optics. Then a lot of other people…[in], so called “squeezed states.” There has been a continuous stream of work in this field and most the work on “Squeezed States.”
Yariv:[refers to question] So we talked about noise, we are talking about Louisell, Siegman. OK, Cook, Lee, and Kogelnik. [question 3]
Bromberg:That was just taken from looking at what you were publishing.
Yariv:Cook and Lee are people that I interacted with during this early period of Esaki diodes research. Cook was a collaborator and Lee and I got involved momentarily when there was a paper that claimed that noise in tunnel diodes could be less than it should be. It had to do again with noise and statistical correlation. In a tunnel diode you have two electron currents tunneling in opposite directions, and the question of the correlation between them. So those are some fundamental issues of noise. Cook was a microwave experimenter who worked with me in the microwave stage, and then left and became a department head of another group at Bell Labs. Kogelnik appeared on the scene at that point—another one of Rudy Kompfner’s recruits.
Bromberg:Oh, he was just coming in then?
Yariv:Yes, about 1960, maybe a year after I joined Bell. Rudy Kompfner used to go to Europe every year and, there are some jokes that he held recruiting sessions in some basement at Oxford, and came back with unusual Europeans with odd backgrounds. Kogelnik was one of his finds obviously he had good taste. He did very well. So Kogelnik showed up at Bell, and immediately…I guess his first efforts were in working with Boyd and Jim Gordon on the theory of confocal resonators in optical beams, and really made some absolutely first, rate contributions, including the ABCD formalism of Gaussian beams. He was, in a my opinion, at the point probably the best electromagnetic theoretician of lasers certainly the most elegant. The reason that Kogelnik and I got together on a paper, which is somewhere maybe on your list, is a noise analysis of optical laser amplifiers. Kogelnik and Yariv, Considerations of noise and Schemes for its Reduction in Laser Amplifiers.” Proc IEEE Vol. 52 (1964) What we did here is essentially look… Suppose we take a laser and remove the mirrors, so now we have an amplifier. Light can come in from one side, exit from the other and in the process get amplified. How good is that amplifier? What is its noise performance? You know, one characterizes amplifiers by a noise figure or a noise temperature. Is it as good as the theoretical limit that has been predicted for masers or other devices? But when you analyze a laser amplifier, you now have to consider more factors. You have to consider the optics, how many modes does it accept and amplify? Something which today we call “spatial filtering”, because otherwise you don’t get the limiting noise figure. You have to make sure that the amplifier only sees the mode that you want to amplify. If it is open to more radiation, it will amplify the noise and it won’t be as noiseless as it can be. So, to obtain the limiting noise figure of a laser amplifier, you have to include aspects of optical beams and resonators. So this is the union between Kogelnik and myself here.
Yariv:I brought in the laser physics background, Kogelnik brought his understanding of optical beams, and together we were able really to show that if you do things right you have an ideal amplifier. Ideal in the sense that it agrees with the quantum mechanical limitations. I like that paper, I think of it again as a basic paper.
Bromberg:Good. I just looked at the first paragraph I think. All I remember from it is your saying that other people deal with noise mode by mode, but in practical schemes you don’t usually get that kind of separation, so you can’t really think in terms of individual modes. That was about all I looked at. But there is a lot more in it than I saw.
Yariv:Yes. It turns out the people who work in microwaves—life is easy for them, they have a microwave cavity that resonates in a single mode. But if you take a laser, all of a sudden, what are the modes? It turns out you have to think also in a way in modes, but the modes are much more subtle now. And you have to do the optics right, that you indeed are left with what is equivalent to a single mode, and the recipes for doing it, and the spirit are really spelled in that paper.
Bromberg:Good. I should like to have us single out in this way the major papers, because in this large bibliography, you know, at first glance I didn’t have a good feeling for…
Yariv:Well, I guess we touched on…
Bromberg:So number 9 is of course, and number 30.
Yariv:Number 9? Number 5 was, in a way. You talked about the non-linear aspects… OK, there are papers which were not important in a sense that they were starting points for disciplines, or for areas of research, but some of the early maser papers…Paper 22…Oh, yes, papers 20 and 21.
Yariv:In a way they are the basis for much of the work later on in integrated optics, but we will come back to them. I consider them important. And 24, which is the experimental, also part of the same package. And then 22 is the first review paper on lasers. And not important in breaking any new grounds, but only important in the sense it probably was used as the textbook in those early days when there were none. A lot of people, schools used it. It is a review paper that Jim Gordon and I wrote, and probably one of the early attempts to look at quantum electronics as a discipline, and to make a theoretical body out of the whole field.
Bromberg:Now, one thing I want to ask you is, as you started to go into this work on calcium tungstate, on these solid state masers, you say ‘60 that began, I wonder what the management positions were. Now management for you would be Kompfner I guess I don’t know who else would be people who were in management. Would this be a matter of people coming down the hall and saying, “Gee, we would like to get more work in this at this time?” Or, how did these kind of decisions get made?
Yariv:I’d say the management in the sense of telling you exactly what project to work on scientific direction didn’t exist. We were really a bunch of people that could do first about whatever they wanted to as long as we did not wonder off into completely strange lands. We were just regrouping, changing directions. I don’t think anybody ever at Bell told me, “Work on this,” or “Work on that,” in the four years I have been there.
Bromberg:So when you started to work on the doped crystal masers, do you remember why you decided at that point? Or why this group got together?
Yariv:Yes. More or less. I decided first to work on optical masers, and so there were other groups at Bell working at trying to make masers from solid state materials—crystals doped with impurities—you needed some doping. My own background was in paramagnetic resonance, where again you work with crystals with impurities, paramagnetic impurities. So I started reading, and as a matter of fact I have a folder here that dates to that period at Bell. Most of the references here are papers on paramagnetic resonance. A lot of papers by the group at Johns Hopkins, of Diecke and Abrams. Gerhard Diecke who with his group had done an enormous work on paramagnetism of a whole group of crystals simply to get acquainted with what can be done. And then deciding that if I was going to work on paramagnetic crystals, you are kind of limited. There were people at Bell growing those crystals, and one of them was Kurt Nassau. So I started working with Kurt, you know. Kurt knew how to grow calcium tungstate crystals so that’s why we started looking at calcium tungstate. And for a crystal grower to learn to grow another class of crystals is very complicated; it might take him a year, two years.
Bromberg:I didn’t know that.
Yariv:It’s not like today you grow one crystal, tomorrow another. You have to get things right, the growth conditions, the thermodynamics, the melt. And so we naturally started working with the crystals that Kurt Nassau knew how to grow, which were calcium tungstate. And he would add different impurities, mostly according to our requests: Fraseodynium, Ytterbium, Erbiu, and we simply tried them all. And one of the early crystals to go was Praseodymium in calcium tungstate probably was the third solid state maser. The first being ruby, maybe the second one was calcium fluoride uranium made by Stevenson and Sorokin at IBM. Praseodymium, in calcium tungstate might have been the third. It’s not too important in retrospect. So really no direction. The grouping was spontaneous.
Bromberg:How did Porto come into it?
Yariv:Porto was hired at that point—was the new kid on the block—and he really didn’t have anything definite to do. Porto had been a student of Diecke, the man whose fundamental work on fluorescence.
Bromberg:I just assumed he came right up from Brazil.
Yariv:No. He didn’t, he had a background. He had graduated from Johns Hopkins, went back to Brazil and worked as an instructor in a Brazilian air force university in a town named San Jose dos Compos, and then for one reason or another—I’m not sure what the reasons were—oh—dissatisfaction, or scientific drive decided to come to Bell. And he was one of the few cases where Bell hired somebody who was already working somewhere else. They usually come straight from college, so that was an exception. And because of his background in spectroscopy he joined me. Of course, it was all spontaneous. It’s possible somebody suggested that he join us, so he first asked me would I like to work with him, and of course I agreed. It turned out to be a fruitful cooperation. Later on, another Brazilian joined us: R.C.C. Leite. That’s on the work on semi-conductors.
Bromberg:Let’s see if there is anybody else on this list that is…I can’t see anyone else in this…Most of the people are perfectly natural, like Boyd, and Collins.
Yariv:I don’t think of the work on the masers there as very exciting, you know, we didn’t really make the first maser, we just made some masers that worked. It was hard work in retrospect. At the time I didn’t realize how many other things I had done turned out to be more important in terms of what I did later on, like the noise work in parametric amplifiers and now we are getting to the work on semiconductors.
Bromberg:Now, but getting away from “in retrospect”, you know, trying to recreate your attitudes as they existed at the time, did it seem important or fruitful? Do you have any recollection of what your estimates might have been at the time of how the work was going? Also, you know, something we ought to get into is the work that was unsuccessful, or unpublished, but is sometimes interesting because it had germs of ideas you could follow up in other ways. So we don’t want to neglect false starts. That’s always an important thing to just keep in mind.
Yariv:Well, much of the work at the time was to take a whole range of crystals with various dopings in them and fashion them into cavities, you know polish their ends, and evaporate silver and make mirrors, and pulse them and make them lase, and look for a material that is really better than anything else.
Bromberg:I see, some will lase and some won’t lase.
Yariv:Some will and some will not. There were crystals like Dysprosium in calcium fluoride doped with uranium which we made go CW, Boyd, myself and Collins. It was a tour de force. We had to cool it and had to use liquid oxygen, which looks almost suicidal today, in order to get rid of the bubbling of liquid nitrogen. And we got it to go, but once we did it, nobody really needed it. It was so difficult to make and the wavelength was 2.5 um or thereabouts. So a lot of hard work. Many laser systems that we got to operate. Praseodymium is not in any great demand or calcium fluoride. Much of that is due to the fact that when another group at Bell, off at left field made Neodymium YAG laser go.
Bromberg:That must have been Johnson’s,
Yariv:No, it wasn’t Johnson: a fellow Geusic, Joe Geusic. He was in the Development department not in the Research department. And Geusic, even before he succeeded in making a laser out of it, already claimed and obviously very competently, that this is going to be a very good system. He studied its fluorescence characteristics, understood it, and when he made it—Neodymium laser—it was so much better than anything else. It had low threshold, worked at room temperature, and it just took wind out of most of the other solid state maser work.
Bromberg:That’s very interesting, because in a way you must have realized as you were making these very difficult experiments that these were not going to be very practical devices.
Yariv:That depends on alternatives. First, we weren’t really thinking about applications so much. At that time I don’t think any of us had a glimpse of the promise of lasers, quantum electronics, yet. And whether they would be useful or not. First, we didn’t think that any laser would be very useful; as useful as they have turned out to be. And then we knew that a lot of other devices were being cooled. The ruby microwave masers were cooled, the parametric amplifiers were cooled to liquid nitrogen temperature to get the noise performance. So, that didn’t bother us so much.
Bromberg:I guess what I’m aiming at, in a way is, given that you weren’t looking towards applications, what was the motivation? Was it science, was it just the sheer joy of seeing how many of these things you could make? What was behind this? Was it a kind of exuberance? It’s the motive I am looking for really.
Yariv:We had a bunch of mostly young people out of universities working on.
Bromberg:We were just talking about the motivations of the group you were working with on solid state optical masers—the questions, problematic—that this work was addressing, to get some feeling for it. You were saying it’s not applications, but it’s science. It was this group of fairly young researchers. Were there certain scientific questions that were being answered?
Yariv:No. I don’t think most of us were motivated primarily by scientific questions. We were motivated—most of us—by the notion that masers are exciting that we were part of a very new direction that they were going to do unusual things. None of us had the glimpse of really how important they were going to be. To pick as an example optical communication. I don’t think that that word was mentioned among us, or that any of us really any of us really had any idea. But we had a vague feeling that it was exciting.
Bromberg:And novel, you could get novel effects?
Yariv:And novel. And also clearly in working with lasers one had to use already then…it brought together disciplines like electromagnetic theory, spectroscopy, quantum mechanics, optics. And it was difficult, and therefore challenging. But by and large most of us were not drawn, at that point, by trying to solve any particular problem except make the lasers. Make them so they work well, and in the process solve the problems to do with them—just a challenge.
Bromberg:I should think, though, that someone like Kompfner would be so much interested in optical communication. Is that a false way of looking at him?
Yariv:When I said that none of us talked much about optical communication. I meant the troops, people like Porto, Kogelnik, myself, who were actually working day to day, Kumar Patel had also joined us by then. Incidentally, we should talk later and come back to this group of people that were in that area then, which was in the retrospect, really unusual. It is possible people with a more global point of view, like Rudy Kompfner, almost for sure must have been thinking ahead, thinking about communication. Because they were in the communication business in a way. We were not. We were just scientists working on it. Our horizons were much more limited. We didn’t have their background experience and we were less aware that we were working for a telephone company.
Yariv:So, that probably, I would say, was the motivation, really the challenge.
Bromberg:It’s nice. It gives a picture of what is going on there a little bit.
Bromberg:The picture of the spirit of the place is really an important part of the story.
Yariv:A lot of people clearly, you realize in retrospect, are very capable, very bright people, with good credentials. P.K. Tien is part of the group and Gary Boyd.
Bromberg:Now Gordon was in the Development area wasn’t he?
Yariv:No. Not E. I. Gordon, but J. P. Gordon.
Yariv:Jim Gordon—the one who made the first microwave ammonia maser as Townes graduate student in Columbia. So you have Tien, Gordon, Boyd, Kogelnik, Kumar Patel. Then later, Porto, myself. I probably missed one or two people. You know most of those people went on later on to really make very important contributions in various aspects of quantum electronics. And those people all at that time were working together, talking to each other every day, and it must have been very exciting. I didn’t realize it at the time, as we were having lunch, that I was sitting people with people who will contribute so many basic aspects of quantum electronics. But clearly, all of them were working hard, certain competitiveness which is kind of inherent in this age period, and the selective process that brings people to Bell, which is a competitive place. So people are working hard, but a lot of interaction.
Bromberg:Competitiveness mostly among yourselves, or…Who were the competitors at that point? Whom did you regard as competitors?...
Yariv:Competition in the sense that everybody likes to feel that he is performing at the level at least comparable that of the people around him. There is always competition with your immediate colleagues, good competition. That is how you judge yourself. You don’t have absolute standards; you always judge yourself relative to your surroundings. And then there was competition with other groups within Bell. As an example, we were in the electronics research lab, and there was the so-called solid state part of the laboratory, where the original work on lasers was done: the group of Schawlow, and Devlin, and Garrett, and Kaiser. Now those people were really bona fide solid state physicists—the background was spectroscopy—and we were by and large electrical engineers, almost the whole group I mentioned.
Bromberg:You’re right, Tien, Patel,...
Yariv:Kogelnik, myself, Boyd. So we in a way had less business in this.
Bromberg:Was there any competition with the Device people?
Yariv:The Device people at that stage were not doing research. Later, as I mentioned, Geusic and company, who made neodymium YAG came from there. And then Gene Gordon and his group, Alan White and Dane Rigden. No there was no competition at that point. Of course, I was not working on gas lasers. There may have been competition between Patel and the people working on gas lasers. I was beginning to veer off at that point and drop away from solid state laser work, and get involved in semiconductor lasers in about 1962.
Bromberg:We should go into that. I don’t know what was going on at Bell in semiconductor lasers. I know what was going on at GE and IBM. I have heard people mutter that Bell was in there even before IBM and GE.
Yariv:That was not true. I don’t think so. I was not in the semiconductor department. There was a semiconductor research group at Bell and there were people working on GaAs p-n junctions. Barry Cohen was one of them and R.J. Archer and J. Whelan, now professor at USC. Up to that point, I had nothing to do with that group. What they were trying to do in that period, I don’t know but I don’t think they were trying to make lasers. I think they were interested in fluorescent properties of the junctions. It was already known they were very efficient fluorescers based on some work, I think from GT&E. To my knowledge there was no attempt to make diode lasers at Bell no real appreciation that it could be done, and the news from GE and IBM of semiconductors lasers made came as a surprise. All of a sudden, I was called to a meeting by a man who was a department head, 2 levels or so up. He had gone to a meeting and in that meeting he had heard about semiconductor lasers. I was there, I don’t remember why. Maybe John Pierce was present. The discussion was, what were these new animals and what should be done about it? I either volunteered or was asked to look into semiconductor masers. At any rate, all of a sudden, I decided to find out about them. I got together this group of semiconductor people whom I hadn’t known before. The group was, first, Roger Leite. He was a new recruit to Bell laboratories, a Brazilian who had gotten his degree from the Sorbonne. The other names, Bond, Cohen, Archer, Whelan were from the semiconductor group. They made the p-n junctions in gallium-arsenide and we made lasers out of them.
Bromberg:You can never tell from the order of names who did what. It can be very mysterious. So, the actual observation that you get this confined.
Yariv:Let us get to confinement. At this point we simply made first efforts to make semiconductor laser and see how they worked. And indeed we got the lasers to lase, at first. And that took…it didn’t take too long, maybe a couple of months, to a point to where we could check and verify. And then of course we, I guess in my case it was myself, started thinking about the theory of those things; you know, why they work.
Yariv:And did some analysis. And was at first surprised that they worked, Incidentally, I wrote a paper recently (it was for some centennial issue of Transactions of the IEEE Professional Group of Microwave Theory and Technique, of the professional group) on the origins of integrated optics. In that I talk a little about this.
Bromberg:I would like to get that.
Yariv:Well put it on your list. I will give it to you later.
Yariv:So I worked out the theory for semiconductor lasers and it became very clear that semiconductor lasers had no business lasing. That the theoretical expression that I had derived predicted the threshold would be so high that you couldn’t really make the laser lase. And fundamentally the reason was the following: it was observed, experimentally, that the radiation coming out of the laser was…the spot was very small if you photographed it or looked at it you say a very narrow stripe 1 or 2 microns high. That means when the light is reflected back into the laser from the mirror it diffracts, because the diffraction angle is inversely proportional to the spot size. And if the light diffracts it means that only a very small fraction is fed back—the rest is being lost. Big loss requires big gain, big pumping. And you couldn’t see how the current could really provide that kind of gain. There wasn’t sufficient enough current, or else the thing will melt. So it was clear that the laser was operating with much smaller loss than that we predicted on the basis of diffraction. And somehow, I forget the exact reasoning, I was lead backward to postulate there had to be dielectric wave guiding going on there. Something was trapping the light like water in a pipe so it could not diffract. The light that came back wasn’t diffracting clearly, as it was bouncing back and forth between the facets.
Bromberg:Now, as you say, it sounds to me like a rather surprising thing to realize, and a rather exciting one. Is that the way it went at the time?
Yariv:It wasn’t exciting, because the excitement often comes in retrospect when you realize how important it is. At the time probably a lot of other ideas were kicked around. And you see, having had a microwave background… I’m sure that the physicists probably even today, certainly then, never heard of a dielectric wave guide, you know the fact that you can form a dielectric wave guide in a pure dielectric, or use the term, although he will understand the principle. But you can find it in boxes like, at that time, Collin’s textbook on microwaves had a chapter.
Bromberg:So it was familiar to you?
Yariv:I knew the animal existed. Also in the electromagnetic theory you know that the presence of electrons can change the dielectric constant of the material, it reduces it. If you have a material with free electrons in it, the index of refraction drops. So, we knew there were elements around that could change their index, but we were looking for something to account for the low losses.
Bromberg:What about the very fact that as you began to examine this it seemed somehow that the losses were just too low? Was that a surprising thing? Any “eureka” moments in there?
Yariv:No, I don’t recall a “eureka” moment.
Yariv:I just think wondering about the losses and then among other things, deciding possibly there was a dielectric wave guiding mechanism there. And then going to look for it. Not knowing what can give rise to dielectric wave guiding, but deciding that…And in paper number 20, it’s a theory for dielectric wave guiding in a p-n junction. Claiming the p-n junction can actually provide a dielectric wave guide. And then paper number 21, actually observing it. This is by operating lasers, not as lasers, but below threshold, and simply trying to look at that fraction of the spontaneous emission which is trapped by the wave guiding. So, basically, in those two papers what we really claim is that the reason the p-n junction works as well as it does is due largely to the fact that the light that is emitted is not free to diffract, but is really shuttled back and forth between the two mirrors trapped in a light “pipe” (waveguide). And the origin of the wave guide was not…we were not too clear on its origin. It turns out nobody still is.
Yariv:But, see those were not what we call today “double heterostructure lasers,” lasers that use both gallium-arsenide and gallium-aluminum arsenide.
Bromberg:You just told me it had to do with the fact that you got free carriers there.
Yariv:That was one possible contribution. You know, we basically called the wave guide in there fortuitous. It was not build intentionally of course, because nobody realized at the beginning that it was there or that it was nice to have it. But it was there all right it was definite. The p-n junction lasers would not have worked except for wave guiding. But we were not smart enough to say, if it is that important, can you make a laser in which this wave guiding is now engineered into the laser. That development took place later. That was due to Alferov in Russia and then Hayashi and Panish in Bell Labs, and Kressel at RCA. That is what made possible to get p-n junctions to work CW at room temperature, below their threshold. So you see we did not carry our finding to their next logical step.
Bromberg:You know it sounds to me as if you became much more conscious of the interaction of application and physics later than you were at this period. Is there anything to that hunch?
Yariv:Yes, I was at that point completely not motivated or aware of applications.
Bromberg:Because it seems to me that in your more recent work there is a strong sense of understanding applications.
Yariv:Yes. That came later, gradually.
Bromberg:So maybe we will talk at some point about it.
Yariv:At that point I became very interested in semi-conductors in general and semi-conductor lasers in particular. And worked on semi-conductor lasers probably for the remainder of my stay at Bell. And that takes me probably to 1963 or so, the time when I left Bell.
Bromberg:Now, I had the guess that you had already begun to think about electro-optic modulation at Bell. Is that so?
Yariv:No. Not at all.
Bromberg:So, in that case let me pause for a minute and clear up something that was a little confusing. You didn’t go directly to Cal Tech, you went to Watkins Johnson and to Lockheed, and then to Cal Tech?
Bromberg:Were they just consulting jobs on your way to Cal Tech?
No. I left Bell Labs and went to Watkins-Johnson, as a job. I wanted to come back to California. And although I had a job offer from Cal Tech, I was trained basically as an engineer—up to a certain point I thought of myself as an engineer—I thought maybe industry really is exciting. So I took a job offer from Watkins-Johnson who had already one or two people working on lasers and had made the decision to get involved in it in a big way. I became manager of the laser department which consisted of maybe myself and two more people, with big plans. So I spent half a year, or maybe three quarters of a year at Watkins-Johnson, and that has to do with applications. All of a sudden I am in a different world now, not a research world. And of course, in industry you need contracts. And it turned out the people I interacted with, the potential sponsors at Watkins-Johnson, were people from the government agencies, I think it was the Air Force, were interested in modulation. And so, I remember sitting at Watkins-Johnson and learning the theory of electro-optic effect, and its use for light modulation. And then being asked one day by somebody in the Air Force, whose name I forgot, could you FM modulate laser radiation. So I thought about it and I had the following train of thoughts: FM modulation is varying the frequency of the laser. What determined the frequency of the laser is the length of the laser.
You know that the longitudinal mode frequencies are length dependent. So, if you took a laser and you could change, modulate its physical length, the frequency will adjust itself at any one moment to the new length. And so you tune the frequency. The frequency will track the length variations. But that’s clumsy, because you cannot move a mirror at very high rates. The next thing I said was, suppose you have a crystal inside the cavity, electro-optic crystal, and now by changing the voltage of the electro-optic crystal you change effectively the length of the laser, because you are now changing the index of refraction of the medium. And that is the same. And that should be a mode of frequency tuning a laser. And I wrote a paper.
Bromberg:Was that a particularly hard idea or easy idea to come by?
Yariv:At the time, trivial. It was an idea, an analysis, that is probably a one day type, once you think of it.
Bromberg:Now, that must be 31 here.
Yes, “Electro-optic Frequency Modulation in Optical Resonators.” That simply describes a resonator which has an electro-crystal inside and if you apply a sinusoidal voltage to it so that the “electrical” length is effectively modulated sinusoidally and the laser frequency tracks it. But in the paper, in the last few paragraphs, I ask a question and answer it myself. My original idea here was that you change the length of the laser and the frequency adjusts to the instantaneous. Let’s say you start with a laser of this length; it is full of photons at a given frequency. Now you make it shorter. The new photons’ must have a higher frequency (shorter wavelength radiation). What then happens to the old photons? My first thought was that they have to die off (decay, get lost) and now you have to wait for the new frequency to be born and take over. That means that if you do it too fast, it won’t happen, because there is a finite time depending on the optical resonator Q. That won’t be good; because that decay takes time which means that you could not modulate fast. But then I thought, suppose I actually physically move the mirror? As I move it from the old location to the new location the photons are reflected off the moving mirror, and in the process are Doppler shifted in frequency. And if you think about it, they are Doppler shifted into exactly the new frequency which they ought the possess by virtue of the new length. So you don’t have to wait; you are not getting rid of the old photons, you are converting them parametrically and in the process doing work. So in that paper I realized the process of electro-optic modulation inside a crystal is non-linear and what you are doing is generating new frequencies.
That was important because that led to mode locking, where you do just that. So you can see my first thinking of modulation and its application is kind of in response to the real world—I’m in industry now, I’m talking to a potential customer. He wants something done, he wants frequency modulation of lasers. And I begin to think about what happens if you now put the electro-optical crystals inside resonators. And at that point I read European papers by Giers and Muller, in which they describe a series of experiments in which they took electro-optic crystals, put them inside modulators and modulated them, describe a whole bunch of things. But one thing they say is the following: when we modulate the lasers (and you know my quotation may be inexact) when we modulate them at certain frequencies which correspond to the longitudinal mode spacing, you know to the Fabry-Perot resonance, when they applied frequency to the electro-optic crystal it happens to be equal to the difference in frequency between the two adjacent modes, the laser undergoes violent pulsations and all hell breaks loose. (That’s not their exact language.) And by and large we avoid this region because it is not convenient and we don’t like it. And so that intrigued me and I’m still at Watkins-Johnson and there is no lab and I do mostly theoretical work.
Bromberg:I see. OK.
Yariv:So I solved the theoretical problem of an optical resonator, with an electro-optical crystal inside when the modulation signal is equal to the difference between modes, and I essentially developed the theory of mode locking. But unfortunately I did not call it mode locking. This is paper 40 “Internal Modulation in Multimode Laser Oscillators.”
Bromberg:And this is all still before you hear about Hargrove and Pollack?
Yariv:That’s before I hear about Hargrove and Pollack experiment and then I decided to submit a paper to this tube conference, which that year is held in Cornell. And on the way, I stop at Bell and visit my old buddies; I had left Bell a year or so earlier. Talked to Jim Gordon, He asked what are you doing. I said, well I am going to give a paper at Cornell. The Cornell papers were not…there was no program published. You came and found out what you were going to hear about. And I told him what it is I was going to talk about at Cornell. This is the paper, incidentally, the analysis which essentially predicted all the properties of mode locking, and probably the first to predict the relationship between pulse width and gain linewidth.
Yariv:That meeting was called “Electron Device Research.” The “Tube Conference” was its unofficial name. Jim Gordon heard about it, and of course being very much aware of the work of Pollack, Hargrove, and those at Bell, immediately called a meeting at Bell, a seminar, organized in a couple of hours, and I gave that talk at Bell. It turned out that a bunch of observations to do just with that work were performed so the meeting at Cornell had the paper by Hargrove and Pollack as well as my theoretical paper.
Bromberg:I see. And as I understand this, that was where De Maria first began to find out about the Hargrove and so on work.
Yariv:I think so, yes. I don’t think De Maria is making claims for… I think he learned about mode locking there, or about that time. The work was published.
Yariv:I’m not sure whether he was at Cornell physically.
Bromberg:Well, he told me that he began to think about this whole problem when he heard the Hargrove, Pollack, Fork paper. And he also mentioned you were at that meeting. Then he went back to UT and began to do it in a…
Yariv:OK. That’s interesting. The main idea of that paper was that essentially when you modulate inside a laser, the so called “modes” of the laser are no longer modes. The new modes, the Kosher mode in the presence of a temporally modulated element in the cavity is now a superposition of the old ones, with well defined phase relationships.
Yariv:I myself didn’t foresee the possibility when I started doing it: it was really a theoretical exercise in how do you analyze an optical oscillator when one of its elements is modulated, and because of the Gurs-Muller comments I centered on modulation at the frequency where all hell broke loose for them. But the idea now that you must really take the new modes as super position of the old ones, and the fact that the phases are now locked, drops out as an automatic consequence of the theory. The modes are no longer independent. And the significance of the “magic” frequency—why things happen only when you modulate at that particular frequency, the round trip frequency—is that this is the frequency for which the side bands of each mode fall on top the frequencies of the adjacent modes. So this is when the modes really begin to exchange information, and are forced into mutual interdependent cohabitation. Now this point of view is very different from the point of view that Margrove and Co., had they used the idea of a time periodic gate. It turns out that their intuitive point of view is very useful for invention; I think that one is far less likely to invent a mode locked, ultra short pulse laser following my formal approach. But once I got it, at least I was able to trust the analysis after numerous checks for errors and conclude… It turns out for instance that all the claims about the ultimate line width or pulse width of a mode locked laser as being limited inherently only by the line width of the gain, and so on, are in this paper.
Bromberg:From the interaction with Hargrove and Pollack and so on did you get any new things? I mean, did that change in any way your thinking? Or add to your thinking, was there an impact there that we ought to…
Yariv:I don’t recall. I could not think at the time of much more to do in this area. Probably the next round of, to me, important inputs in the whole field came from the later work of De Maria on passive mode locking, and later from the elegant analysis of Siegman and Kuizenga of mode locking, and the work of Hermann Haus. All those people took different points of view in analyzing mode locking, which were very very useful. But the idea of the “super mode”, so called, of the linear combination of the original Fabres-Perot modes as the new legitimate mode. It’s an interesting idea conceptually, and that came out of this work. A spatial equivalent of the supermade has emerged recently out of our work on semiconductor laser arrays at Cal Tech.
Bromberg:Just to clear up. There is one paper that seems to have been written at Lockheed?
Yariv:Yes, the Lockheed paper. What happened is that after about half a year or so at Watkins-Johnson, I decided that industry is not for me. Not that kind of industry in any case. And I had an offer from Cal Tech, a standing offer that was extended a year or so earlier, so I called my friend at Cal Tech, Roy Gould, and told him that I was ready to come, and I knew Cal Tech, and was ready to come and work there. But there was something I wanted to do before I came to Cal Tech. I realized when I came to Cal Tech it will take me a couple of years or so to really get going experimentally, set up. And while at Watkins-Johnson, as you can see from the two papers we have discussed—the mode locking and the electro-optic modulation—I became interested in the topics of internal modulation and parametric interactions of light, and there was a very simple experiment I thought could be done to observe some of these things, and I fortunately succeeded in convincing the research lab of Lockheed, which is also in Palo Alto, which is where Watkins-Johnson is, to allow me to do such an experiment there. So what I did is resign from Watkins-Johnson, like maybe April or May, after half a year with them, and already accepting a job at Cal Tech in September, but spending the summer, 4 months or so, working with the scientist, Don Peterson (his name appears on the paper) and doing an experiment at Lockheed. It was probably one of the fastest experiments. We succeeded in setting up a lab and in actually observing...What we did is modulate an electro-optic crystal placed beam inside a laser and generate optical side bands, which were separated from the original laser frequency by the frequency of the modulation, which is in a way how mode locking works. In mode locking you do the same thing, but in mode locking the new side bands coincide with new Fabry-Perot modes of the resonator. Anyhow, so… That’s the Lockheed story.
Bromberg:Peterson later came to Cal Tech?
Yariv:No. Peterson stayed. Peterson was really a very good microwave electrical engineer who joined forces with me.
Bromberg:Oh, I got him mixed up with Pearson. Sorry. That’s a whole different...
Yariv:Yes. He stayed at Lockheed. As far as I know he is still there. In a way they hired me for 4 months. I had to go through the whole rigmarole, you know, of being hired, the whole thing.
Bromberg:Now we are coming to Cal Tech, and one of the first questions I asked was whether it’s interesting to talk a little bit about the comparison of Cal Tech with Bell or with Berkeley. Whether there was anything in the environment that is going to have an effect on your work, or the direction of problems you stake out. Or that put you in essentially a new kind of situation. The Watkins-Johnson job suddenly put you in a situation where you were talking to the funders and negotiating with them in a certain sense on what problems they were interested in. Did Cal Tech have any qualities that we ought to explore a little bit?
Yariv:Well, first, Cal Tech was a totally new situation for me. Because now I find myself in an academic institution as a professor, not as a student. So that was new. Coming to Cal Tech, I always knew of Cal Tech as a select, small, very good institution, high standards. Knew some of the people, for instance the person who hired me, Roy Gould, was a person whose work I followed closely in my traveling wave tube days.
Bromberg:I see. He is a plasma physics person.
Yariv:Now. But then, before his plasma days he was in traveling wave tubes.
Yariv:Which was a natural transition for many of the traveling wave scientists where you also deal with electrons and magnetic fields. And I like southern California. Although I was at Berkeley for 8 years I wasn’t completely brainwashed by the northern California bias against Southern California. A lot had to do with the fact that even while a student at Berkeley I spent all my summers in San Diego body-surfing. It’s the Israeli, Mediterranean attitude, or background. I love the sun, I love the ocean, and you don’t really have them in northern California. The ocean is too cold to swim in so, I came here and it was a new situation. And really, nobody again to tell me what to do: like Bell in some ways, but different in that you are really alone. There was another professor here already, Nicholas George Nick who is now at Rochester, working on lasers, doing some work with gas lasers.
Bromberg:That whole sort of coupling that you get at Bell, where you work with group and that group, that just…
Yariv:No. You were really in a way alone here, having to decide what you want to do. There was some interaction with Nick. His students, some of them, were very very capable, so it wasn’t a total vacuum but our interests were quite different but really being alone and having to decide what projects you are going to work on, because now you have to work with students. Also to set up a lab, start teaching, organizing courses. In my case I started teaching for the first time courses in lasers and that was new, because there was no such course and I had to generate all my own material. In a way, my text Quantum Electronics was the result of the teaching. We teach one course a year, I also started teaching solid state physics. And that hadn’t been done before at Cal Tech for some reason. The physics department here was, you know, very heavily involved in elementary particles and quantum electrodynamics and astrophysics; there was no solid state physics. So I started teaching a course in solid state physics for the first time at Cal Tech.
Bromberg:There was no particular pressure on you to do any one kind of research or another, or to build up any kind of expertise...?
Yariv:None whatsoever. Of course, I was hired here because of my background in lasers. It was probably anticipated I would do work in that general area.
Yariv:But beyond that there was no attempt at supervision or direction. The mechanism for doing that does not exist either, Cal Tech is really an assembly of professors doing their own thing. Even today, our department heads have very little influence as far as what research is being done. So, every professor here is really independent, and the only real constraint is that if you are doing experimental work, you have to, by and large like any other university, finance it with your own sources, bring in your own money. So there was no direction, no supervision. I had to decide what to do and the direction I went to was essentially a continuation of the two directions which had played the most important role in the last few years, at that time, and one of them was the problem of internal modulation, call it mode locking, and the other one was semi-conductors, semi-conductor lasers.
Bromberg:How did you set out to get funding then?
Yariv:During my half a year at Watkins-Johnson I had met a number of people. I was then thrust into this need to obtain funding for my work in the industry. I had met a few people in the Air Force, Wright Paterson Air Force Field, who became interested in the things we talked about. Part of my thinking about, you will recall, internal modulation and mode locking was stimulated by interaction with those people. And there was one person in particular gave me a contract at Cal Tech to work on problems involved in mode locking and internal modulation, parametric effects inside lasers. So that was ok the first. And then Professor Roy Gould, we talked about him, he is still a professor here in plasma physics, had an ongoing contract for doing physical electronics, and he was able to convince his contract monitor Dr. Arnold Shostar at ONR to let me have part of it, to increase it somewhat and to have some of it go to my new work. And with that money I decided to work on semi-conductor opto-electronics, so that was the beginning of kind of two independent directions. With the difference that now I started working with students.
Bromberg:Yes, now how did that go? What difference does that make?
Yariv:Oh, it makes a very large difference, because once you reach a certain level of student activity—let’s say 4 or 5 students—and they have experiments going on and the fact that at the university you have other responsibilities like teaching, and some committee work. You gradually begin to get involved less and less in turning knobs and doing experimental work (up to that point I was involved directly in the experimental work). Because you simply don’t have the time and you can’t really do justice to the students, because to solve experimental problems you have to be in the lab day and night. If something doesn’t work you have to stick with it for one or two days until you solve a particular problem, and you can’t even mention it in an article. When it’s all over it may be trivial and part of your own education in a way. But, in order to support work with students you have to understand what they are doing; you have to understand the background of what they are doing, the other work, plus you have to invest time in thinking and analyzing and working with them. And the net result is that you really find yourself becoming a theoretician, and after a while, also it forces you to think more globally.
Bromberg:You are a theoretician; you probably are also a strategist.
Yariv:Strategist. You have to look ahead a bit and see how does the work relate to other work; where is it going to take us? So you are beginning to look at the picture both from close range as well as from further apart. It takes you from the real involvement with the actual work, with the joy of seeing it for the first time, but you have the satisfaction, maybe, of seeing the bigger picture, and interacting with more students. At a place like Cal Tech, and other places which are similar, a professor has a tremendous leverage. If you have a bunch of students—capable students—working with you then each student can really take an idea, which might be yours or his or both, and carry it on, and do a lot with it, if the students are capable, and some of our students are exceptionally capable. And they work very hard, at this stage of life and some of them are inventive and productive. The result is that the professor can now work on 5 or 6 topics and push them as fast as you could if you worked yourself, or faster. Some of the students are brighter than you. So, to me that turned out to be the most exhilarating thing about Cal Tech. This simply really being able to now see 3 or 4 things being pursued and nurtured.
Bromberg:I wonder if it has an effect on the kind of creativity one has, because maybe you have one kind of creativity if you are working in the lab on the nitty-gritty part of the experiment, and maybe you had another kind of creativity if you are working this more global way. It’s a kind of vague question; I don’t know if it means anything. But if you look at one’s insights—the novelties one comes up with...
Yariv:I think it begins to push you...it forces you to think now, I use the term more “globally”, and see where things may fit on a broader picture, and probably two of the things that resulted due to that period of my work—the work on integrated opto-electronics and later on on phase conjugation—have to do now with this bigger picture.
Bromberg:It wouldn’t hurt, I don’t think, to talk a little bit about the origins of the integrated optics work, from that point of view. Maybe we will want to drop back, because we really left out a lot about saturable absorbers and stuff like that that we might want to go back to.
Yariv:I think the saturable absorbers. I don’t consider that as an area where we made any real large impact on the field. If you want to go back, it’s fine.
Bromberg:Since we are talking about this anyway, let’s talk a little about, say the integrated optics work and the way the ideas came, the thinking.
Yariv:As I mentioned earlier, probably most of what I will be saying on that topic you can find in this article, called the “Origins or History of Integrated Opto-electronic Circuits,” something like that which was published...
Bromberg:I would like you to consciously keep away from what you have already written, because the whole purpose of these interviews is to fill gaps in the written record. There is no point in rehearsing something that is already on paper.
Yariv:OK. So we won’t go through a chronology of what...But maybe just a few comments that I am not sure that I included in that article, that, as I mentioned, when I came to Cal Tech the two projects were semi-conductor opto-electronics and the other one was mode locking, internal modulation. And at Bell, as you recall, the last thing I did in semi-conductors was work on guiding.
Yariv:So, one of my very first students here was put to work on the subject area of guiding in semi-conductors.
Bromberg:That was who?
Bromberg:Ok, so that already brings us to about 1970.
Yariv:I came to Cal Tech in 1964 and David Hall probably started working for me in 1966.
Bromberg:OK. I am looking at paper 66 now.
Yariv:Yes, it took a number of years. That was a long haul. I came here in September ‘64, David joined me at about ‘65 as a graduate student, or ‘66, and worked for about 3 years before the first results came. Dave worked on an optical wave guiding in semi-conductors. And again the idea was I felt that the whole topic of guiding was interesting and that we didn’t explore it very well at Bell, we just touched on it, just recognizing that it was there. I wondered if one could do more with it? Nothing definite; no great visions for integrated optics, just interested in a specific topic now, rather narrow at the time—guiding in semi-conductors. And the only difference is instead of looking for guiding in a p-n junction, we decided to look for it on wave guides fabricated on the surface of the semi-conductor because it would be more convenient. We wanted to fabricate the guides by impurity diffusions which will raise the index. So David Hall started working on wave guiding on wave guiding, and about that time, oh about 1968, or ‘69 (I forget the exact date), the first paper on integrated optics started coming up. The paper by Harris and Shubert from U. of Washington and Stu Miller from Bell. And we simply, all of a sudden, realized that Gallium-arsenide made it possible to electronics to “Integrated” Optics. Gallium arsenide had an advantage that none of the other materials considered there possessed, that in gallium-arsenide you could not only make wave guides, and that’s what we were doing actively at the time, we had just done it. But you could also make lasers as had been demonstrated by my Cal Tech colleague Corner Mead, transistors, and therefore you now had the basis of a new technology of incorporating optics, electronics, switching, modulation, guiding, on the same chip. So, this was the beginning of integrated optoelectronics and you can see how it really started with the early work at Bell on guiding and the Cal Tech work of Dave Hall. And the rest of the details you will find in that article that we mentioned.
Bromberg:Now, there are a couple of questions here that I have asked you in the written questions. One is on the Journal of Quantum Electronics. I’m already on page 3, at some point we should go back and see what we haven’t discussed, but I am looking at 3 9 now, and 10 and 11 on that page, 3 through 13. There are a whole bunch of questions.
Yariv:(reading) “Please give us your memories of the founding of the Journal of Quantum Electronics and its early years.”
Bromberg:You see, one is on quantum electronics, one is on consulting whether that had any important effect on your work. One is on government work, and so on.
Yariv:OK, well. Quantum electronics, what are my early recollections? All I recall, that the initiative for the paper came from elsewhere and I was not part of that, I was just asked to join the first board of associate editors. I think the first board of associate editors. I think the first editor was Glen Wade.
Bromberg:Kingston and Glen Wade were editors and you and Gene Gordon were…
Yariv:The associate editors. So that’s all I really recall, nothing momentous, it seemed like a good idea. I was pleased to join, to be asked to help.
Bromberg:So you really were in there to review papers and to solicit papers in a way, to see that…
Yariv:Yes, to help solicit papers, to help choose the reviewers. We were really, in the actual work, would get papers submitted to the magazine and decide who should review it, send it to the reviewer, handle all the correspondence with him and you know…papers usually go through a few iterations. And eventually decide whether the paper is accepted or rejected and then send it on to Glen Wade or Kingston for final inclusion. One thing I recall from that first issue…when the first issue appeared, just before it appeared, as we were talking about it I was in the process of reading to one of my daughters at night, “Gulliver’s Travels,” and in “Gulliver’s Travels” there is a visit to Lagado, one of the books, you know after the visit to Lilliput and Brobdingnag he goes to visit a place named Lagado, and in Lagado he visits a research lab, he calls it “Academy” and there are about 500 scientists, which Jonathan Swift or Gulliver calls “projectors” working in different rooms, and he goes to visit them, He describes some of the experiments that are carried on. And one of them is a scientist who is trying to extract a light beam out of cucumbers, which are sealed in a glass vial.
Bromberg:That’s very nice.
Yariv:So, I took that passage and it was included, you will remember, on the front page on volume I on the first issue of Quantum Electronics.
Bromberg:No I don’t remember.
Yariv:Jonathan Swift didn’t think too much of this experiment, because the whole thing is a satire on scientists. The experiment which he describes next is very unflattering to science. OK, so I think that’s all I have to say about that, “Please discuss your main professional affiliations. Have you been closely tied to and active in any other societies, or relatively independent of them?
Bromberg:Well, it is really just a way to get some feeling for what part of the community is very much active in society affairs, and what might not. To see how the professional societies of IEEE, or OSA or whatever…What it means in terms of the actual laser practitioners.
Yariv:Yes, I have been involved very strongly with the two organizations which run most of the meetings and transact most of the business of quantum electronics. And one is the Optical Society and the other one is the IEEE.
Bromberg:When did you go into the Optical Society?
Yariv:I joined the Optical Society later. How much later, I don’t recall; it might have been as late as 1970, or the early ‘70s.
Bromberg:Oh, that’s quite late.
Yariv:Yes. Because I started through the IEEE and my loyalty, so called, was with IEEE. But later on as the Optical Society became more involved with lasers, it was a natural. Many of the meetings were now jointly sponsored, and I started becoming de facto part of that group, so at some point I joined it officially, and am now strongly involved with both of them.
Bromberg:The IEEE, you must have been in the Electron Devices to start with?
Yariv:Starting in Electron Devices, and then of course when quantum electronics activities became important, I joined those. So that’s the story. I was involved in the meetings, you know, both in CLEA at the time and IQEC. As a matter of fact I had something to do with the beginning of CLEA; I was among the small group of people who decided we need CLEA.
Bromberg:Well, I would like to ask you about that because I’d like to get some historical material on the evolution of the institutions of lasers as well as the science. So why don’t you tell me what you remember about why CLEA was started and what people were thinking about and that kind of thing.
Yariv:CLEA was started by a very small group of people who were then active in running some of the meetings of the IEEE, of quantum electronics, and the publications. I forgot their names exactly: there was a fellow who was in charge of Ad. Com., there was an administrative committee which was more or less responsible for quantum electronic meetings at that time, and he was an engineer from a microwave company, I can get the name—think about it later on. And myself, I was like the IEEE representative on that committee. Gene Gordon was a member. And I seem to remember at the time, a bunch of us, the three of us, and maybe one or two other people deciding that there was a need for a meeting on the applied aspects of quantum electronics.
Bromberg:That must have been the committee that went between the Microwave Theory and Techniques, and the Electron Devices to oversee the journal.
Yariv:True. And we just proceeded on to organize a meeting think Gene Gordon, if I am not mistaken, was the chairman of the first meeting. And I think that was my contribution to CLEA and I did not remain involved in CLEA affairs after that, although I had served on one or two of their program committees. Because most of my involvement was with the IQEC; I considered myself at the time working really on non-applied topics.
Bromberg:So it really was as a counterpart to the IQEC?
Yariv:Complimentary counterpart. IQEC, it was felt, was not meant to take care of papers in applied aspects of quantum electronics, which clearly were increasing in importance and number. And that we felt that IQEC should remain, “pure,” and that really has stayed the case more or less since then. So that is how this new meeting came about. Nothing momentous.
Yariv:I was very active in IQEC; I was the chairman of one of the Boston meetings, and program committee chairman of the Atlanta meeting. And actively involved, during the last few years, in program committees for IQEC, for CLEA, now CLEO, and also involved in meetings, organizing then, which have to do with integrated optics, and optical communication. Those are mostly OSA type, and OSA organized, optical fiber communication, the integrated optics meeting, and a new meeting which is taking place this year on ultra-short pulses.
Bromberg:That’s just a single topical meeting?
Yariv:Topical meeting, yes. So I have been associated with professional organizations in various capacities through the years, continuously. But I probably was never as deeply involved as some people are. I kind of paid my dues, was a chairman of a conference. “Consulting positions: Number 11.” Are we ready for it?
Yariv:Most of my consulting over the early period, especially almost until recently, was for probably two; one is the Hughes Research Laboratories in Malibu where I had been a consultant from 1965 till this year—so for about 20 years. I just quit, just recently. And the other group is the government, mostly DARPA. Now the consulting with Hughes has really been very very important, very fruitful.
Yariv:And…Mostly because the research lab has traditionally been a good place with a lot of very creative, good scientists, and interesting work that goes on there. You are really triggered. And it was mutual. To pick one topic, probably I would say that the whole renaissance, the whole flourishing of phase conjugated optics is due to probably the interaction of Bob Hellwarth with myself and the Hughes Research Lab people. As part of the consulting. That’s how it started.
Bromberg:How does that go? Was it a problem that they posed to you that lay at the base of it, or...?
Yariv:OK, we can talk about it, take a little while. My story starts at Cal Tech, with teaching—-and we talked about how teaching impacts the researcher? Teaching is also interesting because it forces you to dig and go into fundamental issues, even though you might resent it at times, it is often a difficult process. Sometimes you don’t have the time and you have to scrape a lot of rust from your analytical tools, by the net result—like medicine, is good. And so, I was talking in class, in a course on quantum electronics about the problem of dispersion in fibers, on wave guides. And one example of this dispersion is the fact that if you transmit a picture in a fiber, a complicated picture, if you project it onto a multi-mode fiber, the picture gets washed out with distance. And it doesn’t get washed out because of imperfections in the fiber, but because of the inherent electromagnetic properties of wave guides. And there is no wave guide in the world that can overcome this problem. And the reason that you lose the picture in a way, is because when you focus a picture on to a fiber, the input of the fiber, you excite many modes of the fibers, and the modes have different phase velocities, so the picture elements, so to speak, or the modes get out of phase and the picture gets washed out. That’s very much like Hahn’s spin echo experiments—the mathematics and the concepts. The individual spins have different rotations rates. Here the modes which take the place of spins have different propagation velocities. And they get out of step; and you lose the picture. But it is not a chaotic process. It’s deterministic. And when you get the picture on the receiving side, it still is there, except your eye doesn’t recognize it—it has been coded into a strange format. But all you need to do is restore phases. And you have to restore them in a very particular fashion; you have to take each phase accumulated by each mode in transit and take it away. You have to subtract that addition. That gets you to the heart of complex conjugation in mathematics, which inverts phases.
Bromberg:Now is all of this that you are describing to me stuff that you are beginning to clarify for these lectures? Or is this a mixture?
Yariv:Well, it happens, yes, in my lectures at Cal Tech. I pose to my class the question of how can we transmit a picture in fibers, why can’t we transmit it, first? The whole idea of expressing the picture in the fiber as a superposition of fiber modes is trivial, but the first time it was done—that’s a little insight right there, to realize that’s how you should look at the problem in a multi-mode fiber and it’s trivial, absolutely, but it was a new revelation because it clarified the mechanism for the loss of the picture. So I posed the question to the class, and then pointed out to them that there was a mathematical solution to the problem. If we could complex conjugate the wave, then we can restore the picture. Then we can transmit pictures over fiber, miles, let’s say.
Bromberg:By the way, you said there was an insight here, that was your insight?
Yariv:Yes, that was mine. I published the idea in 1976.
Yariv:But the solution was mathematical, so the next question was how does one perform complex conjugation on an optical wave? How do you, in real time get a complex conjugate replica of an optical beam. In parallel you know, I consulted Hughes; I go there every Friday and at Hughes a scientist-whose name I again forgot, but later on, tomorrow, next week I’ll give you all the names, when you…has dug up a Russian paper written in ‘71 or ‘72 in which a scientist has observed that when a laser beam enters a medium and causes stimulated Brillouin scattering (SBS) the scattered beam coming back is sometimes sharper than the beam that got in. And even if you add an aberration in that path, the beam that came back is non aberrated in it. Certain phase healing, so to speak. And he had been aware of that and pointed out that article to his management and to us—by us I mean Bob Hellwarth and myself.
Bromberg:This was in response to your talking about this, or?
Yariv:No, Absolutely independent. And what happened at Hughes was that we were all at first skeptical that that experiment could be true.
Yariv:Also, the Russian—his name is Zeldovich—who explained the thing, his article was very mathematical and using the language almost of somebody not out of the laser fraternity—was not easily understood. So one doubted the whole thing. And all of a sudden, then the next step is, one day I had an idea. And the idea is that one can use nonlinear optics for phase conjugation. Now I can take the picture field that you need to phase conjugate to restore the picture, and if I “mix” the same “mixing” type that goes on in the parametric amplifiers and other nonlinear optical phenomena and so on, the result is a new optical beam which is a complex conjugate of the first one. Now you know, the conjugation was always there in nonlinear optics. All the years when we talk about frequency generation, you know when you have two fields, E1 at W1 and E2 at W2 you want to get W1-W2, that effect is due to a polarization which is proportional to some constant times the product of E and the conjugate times explain [i(W10W2)t]. Because when you want to seibhart frequency as in (W1W2) that comes along with the conjugate of the corresponding field amplitude. So this thing was staring at us. It’s in every book on Quantum Electronics, but we never paid much attention to the complex conjugate, we always concentrated on the frequencies. All of a sudden one day I realized that what it really meant is that the field E2 is not a plane wave, but a more complex field, with a picture in it, that that process gives rise to the complex conjugate of E2 just what we wanted therefore I had a means. So I wrote a paper on the use of mixing, parametric type mixing due to the second order optical nonlinearities, to solve the problem of complex conjugation, which I needed to get in order to restore my picture after exiting the fiber, That was mine, And then all of a sudden I saw the connection between that and the Russian paper. And I went to read back the Zeldovich paper and I saw that, indeed, he used the word “phase conjugation”—he knew that all along. But he did it in a stimulated process like Brillouin, and I proposed to do it in a controlled non-threshold process, so I read his paper and studies very carefully and told the people at Hughes I believe it.
Bromberg:They had no particular problem in mind. This was a kind of curiosity for them?
Yariv:Curiosity. It looked like if what the Russian said works, it might be important.
Correct for laser beam aberrations in big laser resonators. Now, Bob Hellwarth also has been a consultant to Hughes and Bob and I consult on the same days; and we usually sit in the library talking to each other. Bob has been the world’s leading expert on third order mixing—this mixing, when you mix two waves and get a third wave is second order mixing, two waves come in, a third is generated like in frequency addition. The next order mixing, if you go to nonlinearities where three waves come in and a fourth wave coming out you have the third order. And Bob essentially realized that four wave mixing was even a better way of achieving phase conjugation because you regain the initial frequency and you are automatically phase matched. Bob started work at USC; I immediately started working at Cal Tech. What the Hughes people did is then go back and redo the Russian experiments, and quantify it. And I think this is really the beginning of phase conjugate optics, although the germs are there in the Russian work, the original Russian work seemed to have been dropped.
The Russians seemed not to have fully recognized the impact of the work. Also what we have done at Cal Tech was to apply the formalism of coupled modes (remember the beginning when we talked about John Pierce in Berkeley and coupled modes) to optical phase conjugation. This showed right away that you can also get that in the process of phase conjugation you can obtain gain and because of the inherent backward feedback get infinite gain i.e. oscillation. The simple coupled mode equations have become the starting point for most analyses of phase conjugate optics. Another major development that came of this work was the realization that by replacing one of the mirrors of a laser oscillator by a phase conjugate reflector the resulting phase conjugate oscillator (PCO) possesses unique properties. We demonstrated such resonators at Cal Tech in 1979. By we I mean graduate students J. Au Yeung and D. Pepper and a green thumb experimentalist D. Febete who was a post doc from Israel. The most interesting feature of these new lasers was that they could compensate dynamically in real time for optical distortion inside the resonator. I would also like to mention the first demonstration of phase conjugation to real time image processing because this will in time become “big business”. This arose from a simple realization that phase conjugation was essentially (and formally) capable of doing in real time everything holography could do so that we could duplicate in real time using nonlinear optics some of the classical holographic work of Van der Lugt on image processing and image filtering. So anyhow, to go back—in this case I think the interaction between industry and university has been crucial. The thing probably would not have happened except for my consulting at Hughes.
Bromberg:Essentially it is another little universe of ideas and people and interests. I mean, that is what I am picking up from you. Is that correct? They put you in constant touch with…
Yariv:Regardless of how good you might be, you have to be asked questions. And if you sit alone in a room, you are very limited in the questions you can ask. You have to tap into a bigger pool of questions. It is like evolution—there has to be a pool of genes so that the proper combination can survive. There has to be a pool of questions. So, consulting of any kind, you often found that the questions that at the time you did not appreciate, even those you did not think highly of, force you to think along new lines of thought that you normally would not pursue. And in my case I find that I depend very strongly on external stimulation—if I look at some of the things which I think are interesting and doing...The trigger may be always external, for instance, going to schawanga and hearing this thing about...or the phase conjugate, or mode locking, or somebody from a government agency who probably didn’t understand much about that field asking how does one frequency modulate lasers? So you need continuous flux of questions.
Bromberg:Now you said the other big consulting was with DARPA. (Defense Advanced Research Projects Agency).
Yariv:Yes. This is consulting with a group at DARPA called the material research council. It is a group of college professors advising DARPA on material related technologies. This meets once a year in La Jolla for a few weeks working to work on problems to do with materials, and new technologies, you know as they might impact 10-20 years from now. And that’s also very interesting. Because first we are given a lot of freedom and it’s another pool of questions. One of them for instance had to do with how do we use fibers to detect magnetic fields, and a result of that, I came up with the basic invention with H. Winsor of the magnetic fiber sensor.
Bromberg:Now, with for example your Air Force funding or the other contract you have here, do you ever get the same situation you had at Watkins-Johnson that the funders will come to you and say, “You know, we really would like to do a certain thing?”
Yariv:No. Because the funding now comes from the research agencies of those groups.
Yariv:And they usually are not mission oriented in the sense that they have a problem they want to solve and they go around shopping for somebody who will do it for them. Although some of them might have to justify the money they give you in terms of perceived or present needs of the agencies. But it is a less direct one, and they have a great deal of flexibility. So, by and large the money is earmarked for general area, but with a few strings.
Bromberg:Now, [questions] 13 and 14, we have a few minutes of tape left, and I would just like to look at those questions.
Yariv:“Did the presence of firms like Spectra-Physics facilitate your research or affect it anyway?” Well, mostly for the fact that I could buy various lasers made by those companies, but I myself personally had not much interaction with the laser industry until very recently, but not during that period. My only interaction was really with the big aerospace research laboratories, but not with the industry.
Bromberg:So it was just a matter that you wouldn’t have to build equipment.
Yariv:You can buy a gas laser instead of building one.
Bromberg:OK, then the end I asked you how you would summarize your own style. But anyway, whether you have any comments you would like to make on your own way to doing science. Now a lot of this has come out of the conversation to this point, but you may want to comment on that question.
Yariv:I think that most of the stuff, almost all the projects that I worked on have to do with waves and light. So I guess I do like waves. In retrospect you realized that quantum mechanics is also waves, you think of the electrons as waves and some of my very recent work is now doing some work on properties of electrons in solids, and essentially the wave-like aspects dominate again. Almost probably all the projects I worked on, the common denominator is wave-aspects: manipulation of wave, coherence, mixing. Phase conjugate optics is probably the one area where the intuitive understanding, almost seeing how waves interact has been most productive. I like waves. As I mentioned earlier, I am a body surfer, so my sports have to do with waves too.
Bromberg:I just glanced at the preface to Optical Electronics and I noticed there that you separated the laser field into the atomic aspects and I guess I don’t remember exactly how the phrase...
Yariv:Probably the electromagnetic aspect, or...
Bromberg:Yes. So that made me think maybe the textbook itself, Optical Electronics, in some way mirrored your own, the region of the field you had separated off for yourself.
Yariv:Well, I wrote two books. One is Quantum Electronics, where the emphasis is on the quantum mechanics. Optical Electronics is designed for the more applied laser physicist and this includes most of the practitioners of lasers today. Most of the people using the book today probably use the electromagnetic part more than the atomic part. For instance, the whole revolution in the fiber communication: the fibers, the modulation. So both are, you know, the two main bulwarks of the field. It is probably correct to say that on the whole my own leanings and style have to do more with the wave aspect of it. But not to the exclusion of the quantum mechanics. In some areas they are completely interlinked. The work for instance on the non-linear noise analysis of parametric processes with Louis Sell and Siegman is one example where both the modal electromagnetics and the quantum mechanics are interwoven... So are all interactions of light with matter.
Bromberg:There are some people whose papers you feel are much more strongly in the electrical engineering tradition, and others whom you feel are rather strongly in physics tradition. My own sense was that yours are very much in both. That one couldn’t put you towards the engineering or the physics end of the...
Well I have written papers that by themselves would probably look like quantum mechanics, for instance the “Use of Feynman Diagrams in Non-linear Optics,” is one example. Or others were pure quantum mechanics, quantum optics. Others which are very device oriented. I enjoyed immensely the ability to work both in theoretical physics and device engineering. There are problems that cannot be solved without the joint input, and something that we haven’t touched here, recently one of the main activities over the last three or four years here has been noise in semi-conductor lasers—very important topic—and that absolutely required even the injection of three disciplines: quantum optics, electromagnetics, and semi-conductor device theory as well as some tools of stochastic mathematics. This is probably the reason why my interest in lasers has shifted gradually from all kind of lasers to mostly semi-conductors lasers. You need to understand the semi-conductor device aspects, like transistors, and the modal wave aspects (guiding) and the quantum mechanics of the electrons and holes.
A. Yariv, J. R. Singer, J. Kemp, "Radiation Damping Effects in Two Level Maser Theory," Journal of Applied Physics 30 (1959), 265.
A. Yariv and R. Kompfner, "Noise Temperature in Distributed Amplifiers," IRE Transactions on Electron Devices ED-8 (1981), 207-211.
W. H. Louisell, A. Yariv, and A. E. Siegman, in Phys. Review 125 (1961), 1646-1654.
A. Yariv and J. S. Cook, "A Noise Investigation of Tunnel-Diode Microwave Amplifiers" Proc IRE 49 (1961) 739-743; A. Yariv and C. A. Lee "Comment on Shot Noise Smoothing Mechanism..." Ibid. 1695; H. Kogelnik and A. Yariv, "Considerations of Noise and Schemes for its reduction," Ibid. 52 (1964)
W. H. Louisell, A. Yariv, and A. E. Siegman, in Phys. Review 124 (1961), 1646-1654.
A. Yariv and J. S. Cook, "A Noise Investigation of Tunnel-Diode Microwave Amplifiers" Proc IRE 49 (1961) 739-743; A. Yariv adn C. A. Lee "Comment on Shot Noise Smoothing Mechanism..." Ibid. 1695; H. Kogelnik and A. Yariv, "Considerations of Noise and Schemes for its reduction," Ibid 52 (1964)
A. Yariv, "Spontaneous Emission from an Inverted Spin System," J. App. Physics 31 (1960) 740-741. (See p. 4, above).
A. Yariv, R. C. C. Leite, "Dielectric Waveguide Mode of Light Propagation in p-n Junctions," Applied Physics Letters, vol. 2, no. 3, pp. 55-57 (February 1962). W. L. Bond, B. G. Cohen, R. C. C. Leite, A. Yariv, "Observation of the Dielectric-Waveguide Mode of Light Propagation in p-n Junctions." Applied Physics Letters, vol. 2, no. 3, pp. 57-59 (February 1962).
"The Beginning of Integrated Optics" IEEE Trans. on Electron Devices, ED-31, Nov. 1984 (p. 1656)
A. Yariv and R. C. C. Leite, "Dielectric Wave guide Mode of Light Propagation in p-n Junctions," App. Phys. Letters 2 (1962) 55-57.
W. L. Bond, B. G. Cohen, R. C. C. Leite, A. Yariv, "Observation of the Dielectric-Waveguide Mode of Light Propagation in p-n Junctions," Ibid., 57-59.
A. Yariv, "Electro-optic Frequency Modulation in Optical Resonators," Proc. IEEE 52 (1964), 719-720.
J. App. Phys. 36 (1965) 388-391.
D. G. Peterson and A. Yariv, "Detection of Low Angle Optical Scattering by Fabry-Perot Resonance," Applied Optics 5 (1966) 469-470.
D. Hall, A. Yariv, and E. Garmire, "Optical Guiding and Electro-Optic Modulation in GaAs Epitavial Layers," Optics Communication 1 (1970) 403-405.
A. Yariv "On Transmission and Recovery of Time Dimensional Image Information in Optical Waveguides" J. Opt. Soc. Am. Vol. 66, NO. 4, 301 (1976).
A. Yariv "Compensation for Atmospheric Degradation of Optical Beams Transmission by Nonlinear Optical Mixing" Optics Comm. Vol. 21, 49 (1977) Also. Appl. Phys. Lett. Vol. 28, p. 88 (1976).
A. Yariv and D. M. Pepper "Amplified Reflection, Phase conjugation and oscillation in Degenerate Four Wave mixing" Opt. Lett. Vol. 1, p. 16 1977.
Au Yeung, J. D. D. Fekete, D. M. Pepper and A. Yariv "A Theoretical and Experimental investigation of the modes of optical resonators with Phase Conjugate Mirrors". IEEE J. of Quant. Elec. Vol. QE-15 p. 1180, 1979.
D. Pepper, J. Au Yeung, D. Fekete, A. Yariv "Spatial Convolution and Correlation of Optical Fields Via Degenerate Four Wave Mixing" Opt. Lett. Vol. 3, P. 7 (1978) Also J. White A. Yariv. Appl. Phys. Lett. Vol. 37, P. 5 1980.