Notice: We are in the process of migrating Oral History Interview metadata to this new version of our website.
During this migration, the following fields associated with interviews may be incomplete: Institutions, Additional Persons, and Subjects. Our Browse Subjects feature is also affected by this migration.
Please contact [email protected] with any feedback.
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.
In footnotes or endnotes please cite AIP interviews like this:
Interview of Marshall Nathan by Joan Bromberg on 1984 October 17, Niels Bohr Library & Archives, American Institute of Physics, College Park, MD USA, www.aip.org/history-programs/niels-bohr-library/oral-histories/4792
For multiple citations, "AIP" is the preferred abbreviation for the location.
The discovery of the semiconductor laser at IBM in 1962. Competition with other laboratories. Follow-up studies of the use of lasers for optical logic. Managing the research. How technology affects scientific research.
We are going to talk mostly about the injection laser. But I did ask you if you would begin by talking about what you were doing prior to that. I understand that you were working on semiconductors III-V?
Yes. Well, about a year or so before we started working on the laser, I had been working on tunnel diodes, and I stopped doing that and started to work on optical properties, in particular photoluminescence, of III-V compounds, which was really gallium arsenide at the time. That was mostly all we looked at. And also some insulating materials. I started working with Gerry Burns, who had background in insulators, and we did work on ruby and things like that. So that's what I was working on. I never really "undertook" the laser work because I never stopped this other work. It was concurrent with it. But at the time, there was a great deal of interest in the semiconductor laser, as to whether such a thing could be made. It was talked about at meetings, and whenever you saw anybody. It came up often, as to what one could do to make a semiconductor laser, and there was a great deal of groping. One of the first things that called my attention to it was, sometime in the early spring of l962, that Rolf Landauer invited Sumner Mayburg to the Research Center, and he, Mayburg, was working on light-emitting diodes. He gave a talk on them and I remember one thing he said was that, despite the fact that these diodes were in the infra-red, that is, they emitted at 8400 angstroms at 77 degrees Kelvin, you could see them. You could look at the diode and you could see the light.
How did that happen?
Well, you just turn it on and it's sitting there in the Dewar, you know, and you look and there's a little red dot. You could see it. And that indicated that it was intense. And he also said, though he never got much credit for saying it, that the diode was one hundred per cent efficient at turning electrical energy into optical energy. And he based this statement on the fact that the light coming out of the diode was a linear function of the current, which is a somewhat weak but not unreasonable argument. So people here started making light-emitting diodes, in particular Dick Rutz and Rick Dill, who was in Dick's group. One of the first things I did at that time was, I talked with Peter Sorokin. He said, "Well, the way you make a laser is, you take a piece of the material and you put it next to a flash lamp, and then you see if it lases. You make a rectangular parallelepiped out of it and it's going to be a laser." So we did that. I got a piece of gallium arsenide. I guess it was lapped, and we etched it so that it was nice and shiny, and we put it in his Dewar and he turned on the flash lamp and we looked at the radiation. And we saw just the spontaneous emission. We didn't see any stimulated emission. And there was a good reason for that. I didn't understand anything about lasers and he didn't understand anything about gallium arsenide. What was wrong was that the lamp had a time constant of milliseconds, and for the thing to work — I mean, years later people made that kind of gallium arsenide laser — for the thing to work, you needed a lamp with a time constant of nanoseconds. Otherwise the thing would just heat up. You'd never see anything.
That was a kind of joint experiment — you and Sorokin?
Yes, it was a couple of days work. That was in the spring. Then in the summer something else happened. The people at Lincoln Laboratories gave a paper at the IEEE Device Research Conference, I think it was called the Solid State Device Research Conference at the time. It was in June, and they said that they had made gallium arsenide diodes and they had measured the light output of them [with a] thermopile, and found that the diodes were indeed l00 percent efficient. They had a press release and it got into the NEW YORK TIMES. Then, like now, if something gets into the NEW YORK TIMES, it gets the attention of the IBM upper management. So the Research management here were excited, and at that time the effort increased on the diodes.
I guess I don't understand that. If it gets to the management, what does the management come and say to you?
Not the management here, the management of the company. You know, somebody, I don't know who it was, I guess it was Tom Watson, Jr. or somebody in his office, will call up the director of research and say, "Well, what are we doing about that? What is IBM Research doing about that?"
It will go down to the lesser managers?
Yes, and then they'll say, "What have we got going in this area?" And so there was a little bit of fright. As always when something like that happens, we want to look as if we're in a good position.
Was your manager, at this point, Landauer?
No, my manager at this point was Keyes, Bob Keyes.
OK, so he would come over and say, "You're working on this —"
Yes, "What are you doing?" Yes.
And you would want to do more, is that it?
Well, he might try to talk me into doing more, but he didn't. I just remember seeing that thing. I remember Rolf showing me the account in the TIMES. And I know that was in June. I think that's probably when we started making diodes, although I'm not sure, it might have been earlier. I know that suddenly the interest increased at that point.
I saw a theoretical paper by, I think, Dumke in March.
Yes, he is definitely responsible for making an important contribution. These things are all light-emitting diodes. You know, we really didn't make the connection between light-emitting diodes and lasers. As I said, I didn't know anything about lasers until much later. He was responsible for pointing out that if you're going to make a laser in a semiconductor, you'd better do it in something like gallium arsenide. It's not going to work in silicon or germanium. Which, you know, other people had thought it would. Including, I don't know, Ben Lax, or this guy Bernard in France who wrote a paper about it, something about the statistics of the thing. Then at the same time (I didn't know this but I found out later) there was a contract here, a government contract I wasn't aware of it.
I see, this is the one with Fort Monmouth, I guess?
I don't know, I guess this is like many of these kinds of things. When a lot of people are involved, many people end up feeling at various times that they didn't get enough credit, or didn't get the proper credit. That's certainly the case here. Many people were very very upset about this. There were so many people, many many people involved in this thing, Crystal growers, people doing measurements. There was this contract, and I had nothing to do with the contract. I didn't even know it existed. At the time, you know, I didn't want to work on contracts, they sort of tie you down, you have to do reports — it's sort of like having to justify your research. I would feel very differently now, but at that time I felt that I didn't want to be bothered with it.
So really there was a group that was working on this —
No, they weren't really working on it, they just had gotten the contract. Dumke had done this work — again, he wasn't on the contract. The only person who went on the contract that I know of now was Gordon Lasher. There may have been some other people. And he did this work on mode confinement, and he said, "Well, yes, you can make a laser with gallium arsenide, "and one day in the summer, about some time in July, I could probably get the exact date, I went to see him, and I asked him, "OK, we can make a laser, but what do we do? What do we look for?" Because I had no idea. He said, "Well, what happens is, when you get to the hold, the light coming out of it gets very intense. So what you should do is put a detector near a diode, and measure the light output as a function of the input current." And so I did that. I went back to the laboratory and took a couple of diodes and measured the light output as a function of the current.
You were just doing this by yourself at this point?
Yes. And I saw that the light output was a linear function of the current. There was no increase. And I thought, "well, big deal." I measured two or three diodes and nothing happened. So I went back to Gordon and he shrugged his shoulders, and I talked to Bill Dumke, and I said, "Bill, you know, this laser business is nonsense, I don't see anything. Who could be so lucky to find something like that." So Bill thought about it for a while and he came back and said, "Why don't you look at the spectral output?" So I had a spectrometer because we were measuring spectral outputs of things, so I took the diode, with no cavity, put it in front of a spectrometer and put a detector behind it and measured the current, and lo and behold, as I increased the current, the line width output started to get narrower, and I kept increasing the current and it got narrower and narrower until I couldn't measure it any more. I thought, my God, this is it!
Is this something that just takes a day or takes a week or?
No, it takes an hour. I mean, if he told me to do that now, I could do it in an hour, if I had the diodes, which I had at the time. I measured one and it got very very narrow, and I remember what happened was, as I was measuring it, it also was heating up. I kept increasing the current and it would get warm, so that I didn't want that to happen, so I had to keep shortening the pulse width generator. It was rather crude. I kept having to shorten the pulses, and this was what's called a delay line pulse generator, and the pulse width was determined by the length of a piece of coaxial cable, which was a big piece of cable because the losses in it were smaller. I kept taking the wire clippers and cutting the cable down, smaller and smaller and the pulse kept getting narrower and narrower. Anyway, it kept getting sharper and sharper, and as I say, then I took another one and the same thing happened but not as much. But I knew that if the line got to be less than kT, that was it, and it happened in both of them, but one in particular went down to less than an angstrom, a fraction of an angstrom. I couldn't see it any more with the equipment that I had. I was walking on air! So I knew we had observed stimulated emission. I knew that that had occurred.
When was this, about?
This was about October 20 or so, 1962.
— it seems so completely casual. I mean, it's an unexpectedly casual story.
Yes, it was very casual. It really was. You know, the guy that was crucial to the thing as far as I was concerned was Dumke, because I never would have thought of doing that. I didn't know what to measure. He told me. I think Landauer may have encouraged me to do the experiment.
The other thing is that when I saw these five names on the paper, I thought it was a team.
That's an interesting story. OK, how did they get there? Why the five names? That created some hard feelings. I could have just put my own name, as Keyes said I should have done. I never thought of not putting Bill Dumke's name. At the time I was working with Gerry Burns, and we had been doing some other measurements on light-emitting diodes, and we had been working [together], so we were just publishing everything together. We were working in the same laboratory. We were, you know, sort of sharing things, and he started working on it as soon as I saw this. I put his name on the paper, and it was really my decision. Lasher is the guy that I went to initially, and I started the work because of talking to him, so I felt that he deserved to be a co-author. And then Dill was the main guy who was providing the diodes. So that accounts for the five. It really wasn't an organized team of people that we were. And up to this point, there was no real concerted effort. It was just a few days work. We were aware of this problem. Some people have said, well, you know, this is something that management decided they were going to do and they did. From my point of view, it certainly wasn't that. You asked who encouraged it and who discouraged it? Nobody discouraged it. But, you know, I could do whatever I wanted. As I say, I wasn't on any contract or anything. It's possible I knew about it, but it certainly wasn't a big thing in my life at the time that the [contract] existed. And I got stimulation not only from people here but also from people outside. I remember a conversation I had with a close friend of mine, physicist named Doug Warsthauer, "You know, you really ought to do something with that."
You mean with gallium arsenide?
No, with making a laser. He said, "That's the obvious thing, that's a big thing to do if you do something there, that's good." And I know that Rolf (Landauer) certainly was very strongly pushing the thing, but most of the things that he pushed I became aware of later. Like he invited Sumner Mayburg here. He had written a memo saying it was very likely that an injection laser would be discovered within the next year, something like that.
So in a way he created an atmosphere and some aspects of this atmosphere reacted on you, like Mayburg talking about it? He may also have encouraged me to look for line narrowing after Dumke's suggestion.
He deserves credit, I think. He didn't get outside credit. But, there was no reason to make him an author of the paper, no way we could do that. Some people, in particular Dumke, was annoyed at the number of people who did get their names on the paper because it diluted his credit. To me, it didn't make any difference. To me it was actually a positive thing, because, since I was the senior author it gave me strength at the time. I was in a very strong position at the time. What I did after that was, I figured, well, what am I going to do now? And what I decided to do was to get everybody I could to work on semiconductor lasers.
What was your general reasoning for that?
Well, we wrote up the paper. It was a very short paper, two pages long, something like that. And I figured, we have a month when nobody else knows about this. We might as well get everything we can. And the way to get everything you can is to get as many people working on it as possible. And I sort of figured it would be not too hard to do that, because all I had to do was go to someone and say, "Look, here's a chance for you to get something [in this new area]. —" Also there were all kinds of things I wanted to do, and there was just so much to do. There were more things than I could do. It was, you know, a very exciting time. So what I did was, I went around to everybody I knew.
— in other words, you don't just work within your official group?
I didn't work within any official group. I knew a lot of people who had been hired around the same time I was, which was in the past few years, and I just went to them and said, "Why don't you measure this, because we need somebody to do it?" No, Landauer put a couple of people working on it. I remember he put — I'm not sure — Laff? But I went around to several people. If you look in the November 1st issue of the IBM JOURNAL, you'll see there are about seven or eight papers.
I was going to ask about those.
Well, that was as a result of two things. First I went to all these people. My name is on most of those papers. I went to most of those people and told them, "Look, you measure this. [Or] we can make a CW laser." And we did that, and we looked at room temperature lasing and all that kind of thing. We looked to see where the light was coming from, whether it was the p side or the n side, what region of the diode, and I could see that all of these things were just obvious things to do, and I also knew — and there was a lot of self interest — that if I did that, I would be a collaborator. So I didn't have to do very much, you know. I knew where the diodes were. I knew who had the equipment. I used to spend a lot of time sort of talking to people, not doing work in the lab.
In a way you became a kind of informal group leader.
Yes. All of a sudden I was manager of 40 people.
Oh, that's quite a few.
Yes, there were a lot of people involved.
There were only 40 people in the IBM —
Well, maybe it was 12 or 15, but that didn't include the technicians, and also there were some that didn't publish things at that time. But I became sort of the informal leader of all these people, and that annoyed the management here. They got upset because they were losing control of the situation, and Rolf used to write me letters saying, "Look, calm down, it's not so important we do this, not so important we do that —". Actually it turned out that it was rather important, because we didn't know that GE had done this, actually before us, and published it at the same time.
Yes, what happened? I'd be very interested in the details of the reactions when that came out.
Oh boy! Well, we heard about it a couple of days before [publication]. Somehow we got wind of the fact that they had a paper in PHYS REV LETTERS, and that they had the thing. I never saw the paper till after November 1st of that year when they were both published, but when we saw that [it was] sort of like a big downer, you can imagine — I'm sure it was for them too, especially since they had done more. Hall and his co-workers had actually taken a diode and made a resonant cavity, and we hadn't done that. We didn't do that until some time around the middle of October. I'm sorry, wait, I'm off by a month. I said it was done in October. It was really in September when I first saw the laser action. Then we sent the thing in the beginning of October, and it was in the month of October we thought we had the thing to ourselves. [The papers] were all submitted [to the IBM Journal] November 1st because we heard that GE was doing these things so we wanted to get all the results in that we could, before we had to refer to their papers, before we [officially] knew there was any other competition. We knew that they had results of some sort. You know, even after we saw the thing, we didn't know how to make a resonant cavity. It wasn't obvious to us that you should polish the edges of the junction.
We thought maybe you could get the light to go the perpendicular direction, but that turned out to be wrong, because the substrate material was very glossy. Although we should have known because of Gordon Lasher's paper. He had the right answer. But we didn't. He was here, but you know, there wasn't that much coherence to the thing. But then Rutz and Dill came up with the idea of cleaving the ends of the laser. We had no idea how you could polish the edges of these things. But when they came up with this idea of cleaving, then we made these cleaved devices, and first we just cleaved out little squares and put a contact on top, and the light would sort of bounce around in there somehow and you have a cavity. And then we got the idea of making them long and skinny, so that you had a preferred direction, and some of them did show, you know, what I call traditional laser action, where you got a big increase in power. It wasn't until after we saw the GE results that we realized what you have to do is cleave two sides and saw the other sides so that they were rough, so that you really only got light going back and forth in one direction in the plane. You know, we just didn't do that. We subsequently did that and it worked out in some devices. As a matter of fact, the first diode I looked at had these very sharp lines. It was some kind of accidental cavity. I think I looked at 200 more before I saw another similar, one with as much narrowing.
Was that a function of that way the surfaces were made?
Well, I think we were lucky and had some sort of resonant cavity in that. It was a mesa that was etched, and it just happened, you know, that you got these very sharp lines. You typically wouldn't expect that.
How long did it go on with this kind of free form organization?
Oh, it probably persisted as I remember, till the end of November that year, somewhere around two months. Various people worked on it. Then they slowly drifted away and went back to what they were doing. You know, that was that. I continued to work on it through various measurements, for —
Your papers go right into the mid-sixties.
Yes, for several years beyond that. I stopped in the mid-sixties, when I decided I wanted to become a manager. I became a senior manager, and then I started working with younger people and doing various other things, like hot electron effects and things like that.
Until then you were part of the Keyes group?
Yes. I took his job. He went and became a manager of another group and I became manager of his group. That's what happened.
I noticed there was kind of an abrupt point —
— yes, well, that's where I stopped. What happened locally was that the interest initially was, we're going to make some kind of device that will do optical logic, and there was a lot of interest in that. And Alan Fowler did some work on that, and I did some work on it and Gordon Lasher did some theory on that. I think Ian Gunn did some work. We wrote some things up. But pretty soon it became apparent to all of us — particularly Rolf is responsible for this — that optical logic was useless. That all of this stuff that we were doing was interesting physics, but you'd never make a computer out of using optical logic, and the reason for it was that the power is just much too high. You know, you've got to put a lot of devices on a chip, and if you have a laser with a threshold of milliwatts, you know, there's just no way you're going to do that.
It's not completely clear to me how this understanding of the power begins to percolate through. Is it just something that, as you're working, you think "this is ridiculous, the power is going to be too high?"
Yes. That's right. You soon realize that. That's exactly it. Also, —
— was it thought at some point that you could somehow make a device that would have very very small power?
— well, it turns out you can't, because with the laser, when you start to make it small, ... (off tape). It took us a while to appreciate that there was a threshold power. That is, we knew there was a threshold power, but what was the best we could make it? And then, what were transistors doing? There was another problem, too, and this hasn't been solved to this day, that is, that wires are awfully good for carrying signals around. And the signals on them travel very fast. Close to the speed of light. So if you want to take a signal from point A to point B and not get any crosstalk, the wire is a very good thing to use. This was all before anybody had anything to do with optical fibers. That does solve, to a certain extent, that problem. But even then, it's not totally trivial to prevent crosstalk if you make everything very close together.
I guess what I'm really after with that question is to get some feeling for whether a set of experiments were being done in some systematic or unsystematic way, to begin to elucidate how they could be used or not used in optical logic, or whether it just became clear in some way —
— well, we made devices. We made logic devices. We made a memory device, as a matter of fact. We made something that was bistable. That is, it would sit there and it would be either in an on or off state, depending on its past history. It was either lasing or not lasing at the same current. Basically you changed the distribution of current in the device. If the distribution was uniform, it lased. If all the current was concentrated at one point, it didn't lase, because the losses were high. And people have rediscovered that device with more sophistication e.g. quantum wells.
In other words, these things were just not published, this was internal —
— no, this was published. It's in a JOURNAL OF APPLIED PHYSICS paper by me, John Marinace, and Gordon Lasher about 1963 or '64.
What happened that we're not likely to know from looking at the published stuff?
Well, OK, let's see. I think some of the things I've already described. As you've pointed out, and I totally agree, it was a very casual thing. We were very lucky. There was no concerted effort. I was interested in this thing and a few other people were, namely Dumke and Lasher, and Landauer. Secondly, nobody's likely to realize the upheaval that it caused here. All these people started to work on the thing — you know, against their management's wishes. People came here, came to work at this place, with the idea — they still do, I think — that they could sort of do what they want. Even though they have defined research areas, if something else comes along that looks good, they can do it. And that's what happened. That's why I was able to do this. We were actively at that point, discouraged by Landauer and Sol Triebwasser was another guy — I mean, I think they were trying to do their job. Other projects in the lab were important too.
Keyes didn't oppose any of this?
He was my immediate manager, and he was supportive of whatever I did, both before and after this thing happened. He was a very good guy. He's a nice guy. He's the sort of guy who will say "well, why do you want to do that?" As I say, I was measuring photoluminescence, and I guess he was getting flack from Landauer about why we weren't doing more with diodes, which they thought was the important thing, which it was. So he came to me one day and I told him what I was doing. He said, "Why do you want to do that?" you know, and I sort of laughed. I realize now that was probably hard for him to do and it was a good thing to do. So I was getting, as I think about it, some discouragement and encouragement, in the directions that people wanted things to do.
What about in the following stage when you were doing experiments, '63, '64, '65? Is there anything that was going on there or in the whole business of exploring optical logic, that one is not likely to pick up from the publications? Sometimes there are things you do that just don't pan out, so you don't publish them, but they do influence one's thinking.
No, I don't think so. After that it got sort of ordinary. It was like working on any project, working in any area. It wasn't particularly exciting. As a matter of fact, the excitement started to go away fairly quickly. It became a very very crowded field, which is another reason I started to lose interest. Just about every laboratory started to work on semiconductor lasers.
I didn't know that.
Bell Labs, Raytheon, you know — if you look through the literature, you'll see that at that time there were a lot of places that started doing something or other. It wasn't that hard. All you needed was to get gallium arsenide, which you could buy, and make a pn junction, which was fairly easy to do. So the technology was, you know, very very easy, and so a lot of people started working in it. And it looked promising. At that time, remember, a lot of companies were putting money into solid state research, because the transistor had been invented, and they figured that, well, that was just the first of many many devices that were going to revolutionize the electronics industry. But of course — that turned out to be not true. The transistor now is still around, and you know, all these other devices, like the Gunn diode and the tunnel diode didn't become as important.
I didn't know that. I guess by the same token it would seem that this gallium arsenide laser might revolutionize — that they'd better put in an effort —
— well, there was one thing — why was I in III-V compounds?
Yes, that is something I would like to know, why were you?
That started in Poughkeepsie, in 1959. A fellow by the name of Bob Gunther-Mohr decided that what we ought to do was work on high speed gallium arsenide circuits (that we're working on today). Well, III-V compounds in general had very high electron mobility, and therefore they should be good high frequency materials, and what does IBM need for their computers but high frequency fast switching devices. So we started work, working on gallium arsenide transistors, and a lot of people came in. We had all these people. We never would have done it if it hadn't been for all these people growing crystals. We had all this material around.
You came in specifically on that project?
No. No. I didn't start to work on III-V until about 1960. I think it was probably 1962 before I really got into it.
I see, so one should really look at that as kind of part of this whole rather uncertain [situation] — people didn't quite know which of the semiconductor materials was going to be the best?
Yes, and III-V compounds in general, and specifically gallium arsenide because it was the one that was easiest to do technologically, to grow and control and so on, were interesting for high speed devices. At the same time, IBM hired Ian Gunn, and he used to work on hot electron devices, and he started to measure hot electron properties of gallium arsenide, and while he was doing this, he got this annoying oscillation, which turned out to be the Gunn Effect. It was [months] before he realized what that was. It had been predicted, just months earlier by Cyril Hilsum and Ridley and Watkins as to what caused these oscillations. They had predicted it.
Let's see what else is on my list. I don't remember whether we've talked about everything.
OK, [reading] "decision making, opponents and proponents." I said there were plenty of proponents but no opponents. "Institutional context of the work — what resources?" There wasn't any need for any. We had the gallium arsenide in-house and we also had already started making light-emitting diodes, which are spontaneous emitters, and afterwards, the management did increase the project, but informally, for a period of a few months, when the thing was very hot. We got a lot more people working on it. OK, government, yes, we had this government contract.
That was the one with Fort Monmouth?
— yes —
— that somehow then began to fund what you were doing?
No, I never worked on that contract.
I never had anything to do with it. Afterwards I think Lasher continued to work on it. I don't know who else. That's why I said, supported in part by the Signal Corps, in our first publication; that was Gordon who was supported.
Then I assume, when you said money was increased afterwards, that it was kind of regularized.
Yes, the people who just sort of jumped into it, informally, you know, just to get something quick, some of them stayed working on it, but most of them just went back to what they had been doing.
Is a desire to get a publication a very big part of the motivation of this sort of thing?
Yes, Yes, that's what people do. Most physicists, anyway, and we were all physicists at that time, that's primarily what they care about. Also a publication that somebody would read. People have felt that, gee, you know, here's something that's obviously of a lot of interest — it's sort of cream skimming, in a way, because, you know, it's the first of something, and so that's why. I mean that's what motivated people to do this. Clearly whatever you did was going to be important.
Well, you know, there's nobility as well as anything else in that kind of motive.
Yes, that's why it was so easy to get people to do things. You don't have to twist their arm, just say "Here's a piece of the action," so to speak, you know.
I can see that.
And basically that's how scientists are judged. You know. Somebody says, "Let's see his publication list," and then they say, "What did he do?"
It's not just the length, it's the momentousness.
Right, you know, "He worked on this, this paper is important" or something. Publishing now I guess has become much more of a game. Having 200 publications is not unusual for a senior person. They just accumulate — especially for university professors.
I know that well, trying to read through some of those publications.
OK. "The scientific conduct of the work — why experimental and theory investigations were decided upon." Well, there was no decision, but I think I've described how it was done. "Were there any special strengths of the laboratory?" Yes, we had gallium arsenide, and that was very important. And we had these people. All the people were here. Sorokin was here, and while he didn't contribute directly to this, he was clearly interested in this thing. As I said, we tried this earlier experiment. And we had all the optical equipment we needed and pulse generators and stuff like that, so, I think the main problem was we were semiconductor people. We didn't know anything about lasers. We just didn't understand how they worked.
Of course, that meant that you didn't know exactly what you wanted to do, but did that have any other important effects?
No, I think that it just would have been easier to do, you know, we would have got the thing earlier, if we'd understood lasers. [reading] "Competitive ambiance." Well, we had no idea about GE.
Oh, you didn't?
I knew there was a good chance that somebody else was going to do it. I thought it would be Lincoln Labs. It turned out that they were somewhat later. GT and E wasn't a factor. Sumner Mayburg came and visited here, but as far as I know, they never followed up that work. RCA also worked on light-emitting diodes, but again they weren't involved until later. Bell Labs was very funny. There was a colleague of mine there, Dave Thomas, who wrote a paper a few months before this laser paper came out, saying you couldn't make a semiconductor laser.
When I read Hall's article, 1976, I don't know if you saw that issue of —
TRANSACTIONS ON ELECTRON DEVICES?
I contributed a little thing for that. I had at the time the feeling that Rolf Landauer didn't get enough credit, so I gave him a lot of credit in that.
I see. I had the feeling from Hall's paper that they felt very strongly the competitive pressures, and that they wanted to get something to work, you know. It was as if, let's not bother too much to understand it, or let's both go on the parallel tracks of understanding and getting something that we can get out — [I wonder] whether there was anything like that going on here?
Well, I certainly wanted to find this thing, but I had a lot of other experiments going on at the same time. That wasn't my job. That wasn't what I made my job. That wasn't what I thought my job to be — to find the laser. My job was to do physical research on bulk semiconductors. And I was working on things that were closely allied to that, but other things, as well.
Yes. I'm really belaboring something I think you've already explained.
OK. Well. Afterwards, yes, then we felt, "oh boy, we'd better just do this fast!" And they made us hold the thing up, "they" being the patent department, until they could file the patent.
And I remember, they flew the patent down to Washington, which was a big thing in those days. We got a messenger to take the manuscript to APPLIED PHYSICS LETTERS the same day, so we could get the earliest "received" date we could. We knew that there was competitive pressure. Especially since it was so easy. I mean, you didn't have to do anything. You know, if I counted the hours I spent on it — it was sort of like two days of work time.
When that patent thing comes up, is that a frustration? Or do they arrange things so it really —
Well, it might have held off the publication a day or two, something like that. It wasn't a major thing. And because they put the attorneys right on it — you know, they dropped everything they were doing. Bill Dumke spent most or a good part of his time over the next week or two weeks working with the patent attorneys, so they could write a sensible case. IBM's file date preceded all the others. Now the patent has expired so it doesn't matter any more. It never did matter, because you know we're always exchanging patents. But they took that very seriously.
IBM does take patents very seriously. OK — now, "how the decisions were made to patent or publish." It's not a choice publish or patent. We did both. Oh, the other thing that happened — we knew we had competition, and there was a conference going on, the end of October, in Washington, where Lincoln Labs was giving a talk about light-emitting diodes. It was an IEEE Conference, International Electron Device meeting. They still have it every year. And we thought that what they were going to do was announce their laser at this conference. So we submitted a post-deadline paper to that conference. The conference was the last couple of days in October. Once you submitted something to APPLIED PHYSICS LETTERS or PHYS REV LETTERS, you weren't allowed to talk about it until it was published. That was a rule.
So we had a dummy paper in there, a paper about where the light comes from in the gallium arsenide, pn junction, whether it comes from the p side or the n side, and I had submitted this, and it was a post-deadline paper, and I carried with me abstracts of another paper saying that we made a laser. So I went down to that meeting, the day of the talk, in Washington, and if Lincoln Labs said something about a laser, I was going to say, "We made a laser too."
I never heard a story like that about scientific papers.
We were certainly very concerned about other places, in that sense. Very concerned. And GE was a total surprise. We had no idea they'd been working on it. Well, OK "follow up investigations," they persisted for several years, and then it dropped off for a while and re-started again in the early seventies, when people became interested in display and printer and communications applications of lasers. We started another project.
So it dropped off, not only on your part when you became manager, but generally in the lab?
Generally, yes. It almost stopped, and re-stated again in this early seventies, about '73.
That sounds as if you were the principal carrier of this.
No it was started again by Jim McGroddy. If I had maintained my interest, it would have gone on, to a certain extent. It didn't, and when we realized there was no optical logic, and we couldn't think of anything much to do with it for IBM. So nobody was working on it, I think, for two or three years in the lab.
So it wasn't only you. It wasn't just that you stopped.
The laser effort died, but what did maintain itself was an interest in light-emitting diodes. IBM wanted to make displays, and knew that if you had visible LEDs you could make displays, and so we had an effort in making various semiconductors that emitted visible light, throughout that whole period. But we weren't interested particularly in lasers. And as a matter of fact, (Jerry) Woodall and (Hans) Rupprecht discovered or invented the gallium aluminum arsenide light emitting diode, and then they started to work on solar cells, and never pushed it in lasers at all. That work was done by the Russians and Bell Labs and RCA, and they made the heterojunction laser. Without that, you'd have nothing today. There'd be no interest at all in lasers.
The thing that's puzzling me a little bit though is that on the one hand, here all you research people who are going in whatever directions seem scientifically sound to you, whose rewards are publishing physics in some way that the physics community recognizes the momentousness of it —
— yes —
— and yet, the activity dies away because you can't see any application for it.
— well, we're not totally immune to our environment.
But how does the environment actually mediate? How does it get in there?
Well, OK, how does it? — Well, what happens is, if something has an application, an applied project — like the time when this recent laser project was started by McGroddy, laser people decided it could be useful for a particular display involving liquid crystals and for printers and so McGroddy started a group of people making lasers, growing the material and diffusing and making contacts and packaging them and all that sort of stuff. And if you don't have that, it's very hard to do research on your own in this field. That's why technology, especially in an industrial lab, has a big influence on science. I saw something about science influences technology, but certainly the reverse is true also. Maybe even to a greater extent, because you know, that determines really what you can do easily.
You're saying they created a technological base which makes things easy.
Yes, you can then do experiments. Otherwise you can't justify the number of people it takes to do all this business. In order to make lasers and do various measurements on them, it's going to take half a dozen people at the smallest.
That's an interesting point.
And you can't justify that in a scientific project. You just don't get the resources. You couldn't. And that's become more and more true in semiconductor research. In order to do something, most of the work is making the sample. It used to be, you know, we could take a piece of material and in an hour make a sample. Now, it just takes months sometimes to make a sample.
Which would suggest that from this point of view at least, now research would be somewhat more tied in to technological possibilities.
This kind of research, in this field. Yes, I think that's not true of all the fields. I happen to be interested in semiconductor device research, which is totally tied in.
I think that's a very interesting mechanism, which I don't think I've personally thought of. Are there any other mechanisms you might want to add in here, for the way which in one way or another technology does come to have a big input on this presumably autonomous scientific effort?
Well, the other way that technology's important is, when something is technologically important, then you perfect it. You make better and best, and so many times the work that's been done here that's been scientifically important has been done just because something was technologically important and they had the best samples. You couldn't get them any other place. You just could do experiments and see effects that you just never would see otherwise. Particularly in the semiconductor area, a lot of the nice research that's come out of this laboratory has been because of that.
"Early 1960 views" — as I said, optical logic — there was talk of communications, too.
What does IBM care about communications?
Well, there was talk, and there's still talk. Now there's actually a real need for this. You have to communicate between chips in the computer, and people have talked of doing that with lasers. That is very difficult. But certainly then the next thing, chips are mounted on modules and you have to communicate between modules, and modules are mounted on boards, and you have to communicate between boards. And then it becomes practical, and not only practical but useful. And now there's a project going on here which has to do with taking large processers, large computers, and making a network of them, and in order to do that, you need a very very high data rate. You just can't do it with a coaxial cable. The cables are bigger than the computer.