Interview with the 2000 Nobel Laureates in Physics, Zhores I. Alferov, Jack S. Kilby and Herbert Kroemer, 13 December 2000. The interviewer is Joanna Rose, science writer.
The Laureates talk about the development of information technology, what will come after the chip (4:42), the work behind the discoveries (8:44), research in the industry and the university (14:28), and ‘friendly cooperation’ between scientists (22:43).
Zhores Alferov, Jack Kilby and Herbert Kroemer, welcome to the Nobel e-museum and to this conversation, and congratulations to the Nobel Prize in Physics that you received just three days ago. You have changed the world, the world will never be the same again, can we agree on that? Zhores?
Zhores Alferov: No, I just was saying that the second half of our century was really the time when the information technology was developed so much, and /- – -/ just very, very frequently that we have now a post-industrial or informational society. It happened first of all due to one of the great discoveries just after the war, the discovery of the transistor, the invention of transistor. Then of course the discovery of laser-maser principles and of course the work of Jack Kilby, the invention of integrated circuits just absolutely changed the situation in computer science and information technology. I think that Herbert Kroemer and I just added to that the development of new kinds of materials, semi-conductor heterostructures which permit to control absolutely a new way that electron and light fluxes inside of crystals. This gives a lot of new possibilities for high-speed electronics and it was created real optoelectronics by development of heterostructural lasers ILD. But I also should like to mention the name of Nick Holonyak who did very strong contribution to conventional at first p-n junctions visible lasers and pn-junction LEDs.
Herbert Kroemer: Yes, he was the first, we had actually …
Zhores Alferov: Yes, he was the first who created the visible LED and the visible lasers.
Herbert Kroemer: Yes, he’s a good friend, a good friend of both of ours.
Zhores Alferov: Yes.
So would like to have him here?
Herbert Kroemer: Yes, that would be nice.
You are just a few persons but you were highly involved in …
Herbert Kroemer: You see this is actually quite typical of people who get the Nobel Prize that there are a lot of others behind us who do not get similarly honoured but who at least we should occasionally mention their contribution and point it out, so Holonyak certainly in our field is an outstanding example of that.
I would say that we find ourselves in an ever-accelerating circle of smaller and smaller electronics and bigger and bigger economy, and everybody feels like everything is going faster and faster. Do you see an end to all this? Herbert?
Jack Kilby: Perhaps it is going faster but some of these things have taken quite a long time to materialise, the integrated circuit is 40 years old, Dr Kroemer and Alferov’s work is about that old, so when you’re sitting and waiting and watching it, no, it doesn’t seem quite so fast.
Herbert Kroemer: Of course you do not really anticipate what it will lead to, I think the development has not … You cannot predict what the applications will be. One of the reasons it took so long is because it took a while until the full significance of all of our work became clear. If it had become clear right at the beginning … Well, there was no way it could become clear right at the beginning, in fact I was not allowed to work on the double heterostructural laser because …
It was in the 1950s you …
Herbert Kroemer: It was in the early 60s, because I was told it would never have any applications! Zhores was more lucky because he was not in an industrial laboratory.
Zhores Alferov: So I decided to work at it.
Herbert Kroemer: It turns out I had some other interesting topics on which I then worked for the next ten years, but then I returned to the field so there was a period of ten years when I worked on something altogether different, it’s well forgotten now.
If you look at the chip it is at least 40 years old or about 40 years old, so somehow the life expectancy for the chip, maybe It’s like 40 years or 50 ‒ that’s what I mean by coming to the end. What will be after this?
Herbert Kroemer: I certainly get asked this question very often and I frankly refuse to make predictions. One of the reasons I refuse to make predictions, is because I believe the progress in science and technology is an opportunistic one rather than a deterministic one- Particularly when you’re confronted with situations where the discoveries themselves create their own applications, until such time that we can predict the discoveries we cannot predict where it will go, where the applications will go and so I try to be a little bit modest about that one. But I am convinced that things will not come to a stop, I just do not know in which direction it will go.
Do you think that there is a scientific limit for development, for making electronics smaller?
Herbert Kroemer: Ultimately, once you get below the size of the atom you’re in trouble.
Jack Kilby: The very factor I think is the decrease in cost. In 1958 Texas Instruments sold a single, not very good, silicone transistor for 10 dollars, today you can buy several hundred million much better transistors for that price. Nothing else has ever decreased in cost at that rate, so it’s tremendously opened up the field and permitted new applications.
So it’s the applications and the economy which is the driving force behind the development maybe?
Herbert Kroemer: Sometimes it is and sometimes it isn’t, certainly, although I was lead to the heterostructure ideas by an applications question. Once the basic idea had emerged the order turned around, and suddenly there was a basic idea and you were now trying to apply this basic idea to things that were very different from when it got started. I became involved in the heterostructure concept by wanting to make transistors faster and that did in fact happen, but then I realised that the same principles that speed up the transistor could also be applied for light emitting devices and then the flow was from science to application. It’s like a pendulum, it goes back and forth from science to applications, science creates application, applications stimulate new science, and you cannot say it’s one way or the other.
The thing that forced you to make your innovation as I understand was something called the “tyranny of numbers” can you explain that?
Jack Kilby: That phrase came about because people could visualise electronic equipment that would be useful if it could be built, but the existing technology made it too expensive, too bulky and too unreliable just due to the very large number of parts – the tyranny ‒ and it was a very apt phrase for the time.
And this gave you the idea that now I find something else?
Jack Kilby: Yes.
I see, you two worked in parallel, you were in St Petersburg or Leningrad and you were in the US on the same problem?
Zhores Alferov: Initially not knowing of each other. There is, we are working in parallel but it’s also there was some difference because Herbert Kroemer mostly was doing theoretical work and described some principles, makes the proposals. I did also proposal of the double heterostructure lasers and some other devices based on the heterostructures, but I am an experimentalist and first of all I started with the idea of the double heterostructure laser came to my mind, I first of all started to do that, to realise that. I was working at first just with technicians and engineers and the group would grow up, and we realised our ideas first of all, lasers and solar cells and heterostructure /- – -/ and other devices. But I should like to add to the comments of Herbert Kroemer, that It’s really he and I started with the proposals and consideration from the practical point of view, how to improve well known devices like transistor and then laser for instance, but what we did not expect it really, that from this heterostructure research just appear the new region of absolutely fundamental research which we call now low dimensional electro structures. It became possible only due to development of the heterostructures and successes in the technology of new heterostructures, so It’s always connected by this way, the basic fundamental results frequently came out from practical considerations, and vice versa.
Herbert Kroemer: Yes, but because they went beyond the original motivation, that where they created a new application ‒ this pendulum idea is a very good one. In a very real sense this has been true for science and technology for a very, very long time, that one creates the other, it’s a …
Jack Kilby: Although perhaps we are not very many solid-state physicists in the world, almost none before the invention of the transistor, a few thereafter. Today people working on integrated circuit problems are frequently solid-state physicists, so it has expanded the field and I suppose will continue to do so.
Herbert Kroemer: For example, if you look at the meetings of the American Physical Society, the annual meetings, the biggest single meeting is the March meeting which is, well, they now call it condensed matter physics but it’s really solid-state physics, dominated by solid state physics, so this is the biggest single area of activity for physicists as a profession.
Zhores Alferov: Right now if you look for instance for semi-conductor physic international conference which is going every two years, if you look to this conference for instance 20, 30 years ago there were just a few talks about heterostructures. If you look now, it’s practically 2/3 of the conference is the heterostructures and the devices and applications and so on.
Herbert Kroemer: It simply has become a universal ingredient in solid state technology, it’s a tool without which we could no longer do it.
Zhores Alferov: And then solid-state physics also.
Herbert Kroemer: Yes.
Jack Kilby: All the early work on transistors was done by people with very disparate backgrounds: chemists, physicists, electrical engineers. One of the best I knew had a degree in paper making, and all of these people were pulled into the activity and began to develop a common vocabulary and interest.
Zhores Alferov: I knew something about this very important discovery, transistor inventions in 1947, at Bell Telephone. I knew that from John Bardeen, who frequently visited the Ioffe Institute, and I met him the first time in 1960 at the Prague conference and also Nick Holonyak as a first pupil of John. He told me a lot, and what was very interesting in my opinion, that time it was a group created at the Bell Telephone in 1945, Bill Shockley, Walter Brattain, John Bardeen, Gerald Pearson, Gibney and a few technicians. Mervin Kelly, I think, was the vice president of the Bell Telephone, and he formulated as one of the tasks of this group just to check quantum theory for solid state, for condenser for solid state materials. I will be happy if now the vice president of the industrial companies, everywhere, in the United States, in Russia, would be so clever that put the task in an industrial laboratory to check some new principles in physics.
Herbert Kroemer: That was possible under Kelly at Bell Telephone Laboratories, I don’t think It’s possible anymore.
No, so most of the research is going on in the academia so to say?
Herbert Kroemer: I would say in solids, in our field, the research with a capital R has moved from industry to the universities whereas when I came to the United States it was the other way around. I came to the United States in 1954, I then went back once but came again, and I went to an industry laboratory. There was very little going on at the universities …
As a scientist though?
Herbert Kroemer: Yes. There was very little going on at the universities, I think the school at Purdue was probably the most significant part of …
Zhores Alferov: It was the most significant in the United States that time have been done at Bell, at G …
Herbert Kroemer: But I’m talking about specifically a university, there was very little in …
Zhores Alferov: Very little, at IBM …
Herbert Kroemer: It came actually later.
Zhores Alferov: In our country most of the most significant discovery in the area of solid states and in physics in general have been done in the biggest and traditional very strong research centres of the Academy of Science, Ioffe Institute in Leningrad, Lebedev Institute in Moscow, Kurchatov Institute which was just our /- – -/ institute in Moscow. If you look on just the main discoveries which have been done in Russia in physics, just three, these places.
But the institutes as the name of your institute says, is Physical Technical, this is a bridge between science and industry?
Zhores Alferov: Yes, because it’s a bridge between basic research and applications, and of course because I could mention Ioffe, when he founded our institute in 1918 during the civil war time, he was clever enough that physics in the 21stcentury is background for technology, so he from the beginning considered the physical technical institute it was the name of our institute and he was very clever person. He founded simultaneously physical mechanical department of the polytechnical institute. Maybe it was just another example, MIT and later Caltech, the universities where physicists educated with very good understanding in engineering problems and engineers educated with very good basic background in physics and mathematics. These physical technical universities we can call them Caltech, MIT, Physical Technical institute in Moscow is a university, in polytechnical institutes were the places where the new generation of engineers and physicists were trained.
I understand that industry is needed to go ahead with this research too, the devil of applications, so …
Herbert Kroemer: I think the closer you come to the applications the more this becomes something appropriate for industry to do rather than university because many of the problems require solutions reliability, for example requires solutions on a level where you make many, many devices. The reliability problem is probably a key ingredient in leading us from small integrated circuits to large integrated circuits, and that cannot be done at the university.
Jack Kilby: There has been a shift, industry is funding much of the work in the universities on semi-conductors these days, directly through organisations like Semtech, and it’s become significant enough so that the federal government is willing to provide a great deal of support, so industry has chosen to work on more immediate problems.
Herbert Kroemer: I think it’s basically a timescale question. Anything where you know where industry knows this has to be done, but we don’t have to do it this year, we will need it five years from now. That typically then they support university work in that field.
Zhores Alferov: But a thought that just came to my mind right now is that in spite of very heavy situation in Russia from an economic point of view, and in science of course, but there is interesting what I can say, just going interesting prophecies. In the United States the big laboratories in the companies practically do not carry out now basic research. It’s moved mostly to the university, and that in the universities were created scientific centres, like for instance centre of microelectronics and /- – -/ compounds at the University of Illinois, your centre of quantum structures at the Santa Barbara.
Herbert Kroemer: Yes, we just got a new centre on nanostructure and technology.
Zhores Alferov: Nanostructure yes.
Herbert Kroemer: It was just announced last week.
Zhores Alferov: In Russia research will mostly concentrate, the first level of research, not in the university but at the Institution of the Academy of Science. The biggest institutions of the Academy of Science right now develop more and more education systems, and more and more going close to the universities, so this process is good, because it was started by a different way in our country, and very long time ago by Peter the Great. He founded the academy and he founded academy together as research laboratories and the university, and high school gymnasium, so now develops the idea of Peter the Great, but it’s interesting that processes in the United States and processes in Russia going from different direction but …
Herbert Kroemer: … have found the same optimisation to …
Zhores Alferov: Yes.
Herbert Kroemer: That’s called Darwinism.
Jack Kilby: There have been tremendous changes in the universities. I graduated in 1947 as an electrical engineer, and I had never heard of solid-state physics. I should have, the university had a few courses, but it wasn’t considered very important for electrical engineers, and there’s been a tremendous change in that sense.
What I wonder with these couplings to industry is that industry needs secrets, it has to develop their own processing that not everybody knows about, but in science the openness of science is the ethos, I think.
Herbert Kroemer: I think much of the progress in industry, there is something like a pre-competitive stage what we call, this is before you get into production details. There has been a great deal of openness even between these industrial companies, and in particular the Bell Telephone Laboratories where much of this gets started, they were very much interested in seeing, in spreading the information, and they were very much interested in seeing others, even competitors, participate in this when it was too big for them. They realised it was too big to do it alone, and besides, as a monopoly in the telephone system they were not threatened by anybody. By giving away their secrets in the early days, they actually made sure that the overall development would take place more rapidly, but when it comes down to production details then the intellectual property preserves, become very, very strong.
What about your corporation? Was it mostly cooperation or competition?
Herbert Kroemer: I think it was competition, right?
Zhores Alferov: Yes.
Herbert Kroemer: But friendly competition.
Zhores Alferov: It was absolutely friendly competition, and sometimes it was just /- – -/.
Herbert Kroemer: Yes.
Zhores Alferov: I think in general the example of the work in heterostructures, because we had very strong competitions with Bell Telephone Laboratories, with some other laboratories, but it was during the Cold War time, end of 1960s, beginning of 70s, but we invited each other to the laboratories, we organised joint seminars, so it was just an example, I think. The physicists in general gave the example how to live in this complicated world together and enticing the American and Russian physicists during the World War II, during the Cold War ‒ during the World War II we were just on the same front ‒ but during the Cold War also gives the example for friendly competition and to show that we are representatives of science and we are together.
Herbert Kroemer: Yes.
Jack Kilby: I think another important effect was that as the field matured, as work became more expensive, in the early days you could literally build transistors in your garage and all the equipment that was required to do so. Today a modern semiconductor flat costs several billion dollars and it’s desirable for the industrial companies to share the development costs of those things.
Thank you very much for this conversation.
Herbert Kroemer: You’re welcome.
Did you find any typos in this text? We would appreciate your assistance in identifying any errors and to let us know. Thank you for taking the time to report the errors by sending us an e-mail.