Oral-History:Gordon Earl Moore

About Gordon Earl Moore

In the interview, Gordon Earl Moore explains his involvement with the integrated circuit conception at Fairchild. His recollections include who conceptualized and invented the IC and how Fairchild was able to proceed with the ideas of its members. Moore talks of setbacks and troubles which he and his colleagues encountered in the processes of the IC development. In addition, he explains how the IC invention was accomplished in the United States, while it failed to be materialized in England and other European countries. In the interview, he also provides lessons for students of technological innovation. The interview concludes with Moore's explanation on the ways in which Fairchild, though a small company, could reap such a success.

About the Interview

GORDON EARL MOORE: An Interview Conducted by Michael Wolff, IEEE History Center, 4 March 1976

Interview #457 for the IEEE History Center, The Institute of Electrical and Electronics Engineers, Inc.

Copyright Statement

This manuscript is being made available for research purposes only. All literary rights in the manuscript, including the right to publish, are reserved to the IEEE History Center. No part of the manuscript may be quoted for publication without the written permission of the Director of IEEE History Center.

Request for permission to quote for publication should be addressed to the IEEE History Center Oral History Program, 39 Union Street, New Brunswick, NJ 08901-8538 USA. It should include identification of the specific passages to be quoted, anticipated use of the passages, and identification of the user.

It is recommended that this oral history be cited as follows:

Gordon E. Moore, an oral history conducted in 1975 by Michael Wolff, IEEE History Center, New Brunswick, NJ, USA.

Interview

INTERVIEW: Gordon Earl Moore

INTERVIEWER: Michael Wolff

DATE: 4 March 1976

PLACE: [location not given]

Conception of the IC at Fairchild 1959 Meeting

Wolff:

I think 1959 was roughly the conception of the integrated circuit (IC) – at least at Fairchild. How do you recall it?

Moore:

At that time I was head of engineering and Bob (Robert N.) Noyce was director of R&D. We were essentially working in parallel positions. As I recall it, is the original description of the Texas Instruments (TI) part got published in one of the trade journals describing a component that had both the transistors and the resistors in it that did a function with some kind of an oscillator I think. We had different technology developed at Fairchild. Bob assembled a meeting at which I was not present to explore the question, "Given our technology and the idea of putting the several components together, what could we do?" During that meeting he came up with what proved to be two key ideas. One was the idea of using junction isolation to separate the components. He subsequently lost that to Kurt Lehovec in a patent interference. The idea of using an adherent metal film over the silicon oxide was something we had in our planar devices. My view of it is that Jack [St. Clair] Kilby originally had the idea of making an integrated structure, but Fairchild had the technology that made that kind of thing practical. Bob Noyce was the one that came up with the ways of applying the Fairchild technology to achieve the integrated structure.

Wolff:

Do remember just when that meeting occurred?

Moore:

It was probably in early '59.

Wolff:

Do you remember it as being triggered by Kilby's publication?

Moore:

Not necessarily. I think one of these references either from the Air Force people or from Kilby just mentioned that there was such a concept around.

Wolff:

Do you remember who else attended the meeting?

Moore:

I believe it included Jay Last, Jean Hoerni, and possibly Sheldon Roberts. I'm not sure.

Wolff:

Was [Lippincott] Ralls there?

Moore:

I don't know. He may have been. Ralls was our outside patent attorney.

Geoffrey Dummer’s Early Proposals

Wolff:

That's very interesting. Were the people at Fairchild familiar with the ideas of Geoffrey Dummer in England?

Moore:

I was not at that time. I talked with Dummer subsequently.

Wolff:

Did you know that he had proposed such a circuit in the early '50s and that around 1957 the RRE had made a model – the Royal Radar Establishment?

Moore:

I have seen a reference to one of his early statements. In very general terms he defined the kind of thing we were talking about here.

Wolff:

He used the words solid circuit, which is kind of interesting. Do you remember at that time at Fairchild having heard of the term?

Moore:

No, I don't remember. I don't think I knew of Dummer at that time.

Wolff:

It's kind of intriguing to speculate why they didn't invent the thing.

Moore:

I think they were lacking in the technology to do it. Having the technological base is very important in these things coming to pass. This was about the time that the big molecular electronics idea was getting kicked off also by the Air Force. They were looking for things quite contrary to integrated circuits. They were looking for functional blocks in which I think the quartz crystal remains the full useful example – something where you couldn't tell what the individual things were doing but collectively they performed an electronic function. To get his later contracts, Kilby had a tough time selling them on the idea that an integrated circuit was really consistent with what they were after. Was Bob's recollection reasonably close to mine?

Wolff:

There's a difference in a way. I should ask him about this Kilby point. The way he remembers it, is that as a result of the earlier work at Fairchild, particularly the planar technology and metallization – but particularly the planar – Noyce said that Ralls challenged him to think about broadening the planar concept and looking at what else could be done with it. Therefore he started thinking about circuits. He said that one ordinary day it sort of all came together in his head and he sketched on the board his concept of how to put transistors, diodes and resistors together. I guess the idea was isolating multiple domains as a key thing. In his lab notebook from January 1959 he has some sketches, which he says are essentially what he put on the board. He doesn't remember who was in the room. He said maybe you were there, maybe Victor Grinich.

Moore:

Grinich would likely have been there. I'm not sure.

Wolff:

Maybe Ralls. He doesn't remember.

Moore’s Absence from the Meeting

Moore:

Maybe one reason why I remember this really specifically is that it was very soon after we had split the responsibilities there and this happened at a meeting to which I was not invited because I was off doing engineering things rather than worrying about the R&D stuff. I was exposed to the results of it soon after the meeting.

Wolff:

What you have been telling me is what you heard. You weren't in the room.

Moore:

I wasn't in the room. That's right. I heard about it soon after that, and the fact that I wasn't at the meeting probably caused it to leave a more indelible impression on my mind than it would have had I been there.

Wolff:

You could be recalling the same meeting.

Moore:

It could be the same meeting, but I'm quite sure it involved essentially the key guys that were in the R&D group. That group included Last and Hoerni. Grinich would likely have been there. Ralls I would not have known about. He was an outsider that may or may not have been there. My recollection of hearing the description of this meeting was sitting down to say, "Okay, what can we do with our technology and its direction?" Maybe Bob had already thought about this a fair amount of time before that meeting, but he came forth with these ideas of the junction isolation and the interconnection.

Others at the Meeting

Wolff:

That's very interesting. I should get back with him on that one point and ask him, because we never really talked about a motivation and whether he was familiar with the TI publication and whether that was what had done it. Who else would you think I should ask for a recollection of this meeting? [Victor] Grinich?

Moore:

Jay Last might be a good one.

Wolff:

Is he still at Teledyne?

Moore:

No, he's not there anymore. He still lives in Southern California. Grinich would surely be another possibility, but I'm not as sure that Grinich was there.

Wolff:

Grinich is at Stanford. Right?

Moore:

Yes.

Wolff:

How do you think I could find Last? Do you know where he lives?

Moore:

I think he bought Robert Mitchum's house. I think it must be Beverly Hills.

Important Inventions Leading to the IC

Wolff:

How would you describe Bob Noyce's main achievement in this? He's quite a diffident fellow and doesn't blow his own horn. Talking to him, one gets a picture that the rest of you invented a lot of the important things and he just sort of thought of putting them together.

Moore:

<flashmp3>457 - moore - clip 1.mp3</flashmp3>

One gets peculiar views of this. Of the things that we invented in the early days at Fairchild that really made the whole technology go, I'm not even sure that those were the most important – the two that are associated with the integrated circuit – in getting the thing started.

Wolff:

The two being what?

Moore:

The two being the junction isolation and the interconnection, which are usually associated directly with the integrated circuit. I can think of two other things that at least at the time impressed me very much more. One of them was the use of aluminum for the contacts to the transistors. Until that time nobody had any idea at all one metal could be used to make contact with both the emitter and base. I remember struggling with all kinds of metals to try to do this, and one day Bob said, "Why don't you try aluminum?" From everything we knew, aluminum shouldn't work. I said, "Okay. I'll try aluminum." It worked beautifully. That was major breakthrough and got us around a huge hurdle in that technology. There was one other similar thing that is not so obvious with respect to making good transistors that he threw out the same kind of way. These were both cases of proposing something that I really thought wouldn't work when he made the proposal, then running out and trying it and having it get us by significant barriers. Bob was very good at coming up with ideas like that.

Wolff:

The aluminum was important in integrated circuits.

Moore:

It was important in the whole thing. It was important in making a practical double-diffused transistor first of all, which was the first thing we set out to do.

Wolff:

And your recollection related to that.

Moore:

That's right. That really got the double-diffused transistor out of the laboratory as a practical device. A lot of the rest of it was many detailed contributions to making the technology fly. Jean Hoerni is given credit for coming up with the idea of the planar structure and he certainly was its principal champion at Fairchild, but Bob certainly suggested things very close to that prior to Jean's idea on it. It was a pretty close-knit group, and Bob was certainly the idea man in the group. Some of us were detail technologists that made some of these things happen.

Wolff:

Was the double-diffused transistor before the IC?

Moore:

Yes. That was before the planar. We made mesa double-diffused transistors first, and then we went to the planar structure and left the oxide on the junction, which is generally credited to Hoerni. And that was key to having that and aluminum to run the leads up over the aluminum for example. It's hard to say in something that is as evolutionary a technology as this specifically which were the big steps forward that really broke with previous tradition and which were just another significant advance in the evolution. Certainly Bob Noyce's collective contribution through that time period resulted in a major change.

Wolff:

That's interesting. That's very helpful. Do you think the others would agree with that?

Moore:

I think so.

Contributions of Kilby, Noyce and Lehovec

Wolff:

Speaking of contributions, let's look at the actual invention. It is my understanding that the upshot of the whole thing with quartz is that Jack Kilby and Bob Noyce were co-inventors – Bob with the metallization and Kilby with the idea of putting two devices together in a chip.

Moore:

If you take the Patent Office's resolution, it was Kurt Lehovec from the isolation point of view. Bob lost the patent on the junction isolation to Lehovec. All Lehovec ever did was file a paper application with a patent as near as I can tell.

Wolff:

I missed the point.

Moore:

That was the third major invention that was required to make a practical integrated circuit. There was an interference between Noyce and Lehovec on that and the Patent Office gave the patent to Lehovec.

Wolff:

Is Lehovec technically an inventor of the IC?

Moore:

According to the Patent Office. It's one of the important things that was needed. I think in the technical community, because all he did was file a paper patent application, he is not recognized as the inventor. Success has many fathers and all that kind of stuff.

Wolff:

Noyce lost that to him on an interference?

Moore:

Yes.

Wolff:

I didn't know about that. Where was Lehovec?

Moore:

I believe he was at Sprague Electric Company.

Wolff:

The technical community mainly looks at Jack Kilby and Bob Noyce. Is that right?

Moore:

Yes, I think that's right.

Wolff:

Isn't Kilby generally credited with having been the first person to build an IC, though not maybe a practical one?

Moore:

Yes. I credit him with such. In fact, I wrote an article for Encyclopedia Britannica a while back where I said that. They edited it out.

Wolff:

Tell me how you look at it.

Moore:

I see Kilby being the first guy with the idea – and the first structure – of a rudimentary integrated circuit. The Fairchild contribution was essentially the technology that made it practical. Bob Noyce was the key guy in seeing how to tie the Fairchild technology to the idea.

Wolff:

Do you feel Lehovec played a part?

Moore:

In my view, no, he really didn't.

Wolff:

What is Lehovec's patent title?

Moore:

I don't know it's title, but it very directly covers multiple junction isolation.

Wolff:

That's interesting. I didn't know that. You told me when we talked in the fall that you would be helpful in refereeing between Kilby and Bob Noyce. Is that what you meant?

Moore:

Yes, I guess this is really what I meant.

Contrast Between Fairchild and Texas Instrument

Wolff:

That's very helpful. I understand that in 1957 Fairchild's parent company put an 18-month limit on its backing for Fairchild.

Moore:

That was never done as a real limit. That was just kind of a first commitment that they made. Fairchild Semiconductor was formed in 1957 and we put together a plan that was of about 18 months duration, I think $1.4 million, and they said "Okay, fine." However the intent was always that if things were progressing there would be more money available.

Wolff:

Was this in any sense an incentive to push toward this IC?

Moore:

Not really.

Wolff:

Is it true that Fairchild's drive was strictly consumer and with its own money in contrast to Texas Instruments being military?

Moore:

That's right. The Fairchild Semiconductor management felt very strongly that government money was more a hindrance than help. We had no direct government support at all during that time, but our first customers were principally military.

Wolff:

Even though you had not been looking for that?

Moore:

We had not been looking for any direct government support. We were willing to sell them to anybody that wanted them.

Wolff:

You said your first knowledge of the TI work was that trade paper.

Moore:

It was very brief mention of it someplace.

The First Working Model

Wolff:

After this meeting where Bob Noyce promulgated this concept, do you remember what happened after that? Was it tough to go from there to the first working model?

Moore:

I think we put together a project fairly soon after that. Jay Last was project leader on that. I remember at Wescon that fall, which must have been in August, we showed a flip-flop in a TL5 can that wasn't completely the kind of integrated circuit we were hoping to build but which at least showed that much electronics in a small deal.

Wolff:

Was that the 1960 Wescon?

Moore:

It was the 1959 Wescon. However it was a bit of fraud in that it was really kind of a hybrid structure rather than a fully integrated one at that time. I don't remember how much of a hybrid it was except that it had resistors that were made by using one of these IBM electronically readable pencils.

Wolff:

What's that?

Moore:

Those pencils that they used to use for electronically scored tests that were especially conductive. We used those for resistors in the structure on the ceramic and then erased them down to the right value. It was just showing what an integrated circuit would look like more than being a real structure. However we continued on, and we must have had fully integrated structures sometime in 1960. We announced them as product in either late '60 or early '61.

Setbacks

Wolff:

Do you recollect any great setbacks or troubles or did it all go smoothly?

Moore:

It was a tough program. In our very first ones we tried to move away from such things as lead bonding.

Wolff:

What kind?

Moore:

The lead bonding – the sticking of the little leads on them. We cut the integrated circuits out to be the circular little bumps on the side, little grooves, so we could fit them in one of the packages with the leads coming in right around the edges. The leads all came up to the bottom of header and we tried to make a round die with conductors coming out of each of the leads. We put a blob of conductive epoxy in each of those. Assembly took some evolving to come down to the idea of doing the lead bonding in each circuit.

Achieving the isolation was difficult. In order to make the junction isolation, we had to go to very much thinner pieces of silicon and then diffuse completely through from the top and bottom until the diffusion fronts met. That was a process that took considerable doing. It was a much deeper diffusion than anything else we had ever done on silicon devices. The idea of moving this into production grew to be quite formidable. I got very directly involved then and I was in charge of the engineering department still.

Wolff:

Was this about '61 or '62?

Moore:

It was '60 to '61.

Wolff:

I wanted to sort of quit with the first working model.

Moore:

The first working model must have been in late '60. It may have been in '61.

Wolff:

Were these problems you mentioned about getting isolation, the lead bonding and whatnot related to getting that first working model?

Moore:

Certainly getting the isolation was required to get the first one. If I can find my notebook I could probably find when we had our first operating model, but my guess is sometime in 1960.

Historical Significance of the Invention of the IC

Wolff:

Have you ever thought about whether there was any lesson in all of this? Does the invention of the integrated circuit teach us anything about technological innovation in America or California and Texas?

Moore:

I'd hate to generalize.

Wolff:

What might historians of technology find significant in all of this in years from now? Is it possible to imagine or speculate about that?

Moore:

In my experience, the disconcerting thing about something like this is that the people doing it have no idea of the significance at the time they do it. The first family of integrated circuits was completed were very rudimentary things by any reasonable measure. At that I was director of R&D at Fairchild. Sitting down with Jay Last and some of the rest of the people, the conversation was essentially, "Okay, we've done integrated circuits. What will we do next?" as if that was all there really was to it. There was no appreciation at all that this had the potential of being the basis of a revolution in electronics and a billion-dollar industry.

Wolff:

You didn't realize at the time how significant this would be?

Moore:

Absolutely not. Even after a family of integrated circuits was introduced, we didn't have the remotest idea that this was truly a major difference in the way electronics was going to be done in the future.

Wolff:

It got the impression from Bob that he felt they realized what a great market and all that. Of course it is very hard to remember so many years later.

Moore:

I think if he did he was the only one among us who did. We immediately started to look at things like thermoelectrics and one thing and another to see if we could find something else that was going to be fun to do.

Wolff:

When did you first realize how important this would be?

Moore:

I think it was several years later.

Lessons for Students of Innovation

Wolff:

If a historian of science and technology asked you what you thought what kind of lesson was here for students of technological innovation, would you be able to answer that?

Moore:

I think one thing that was very valuable at Fairchild in the work we were able to do was the immaturity of the organization. We did it because it was interesting and exciting. We didn't do it because we understood that there was a huge market out there that we could potentially tap. I think that the same kind of ideas starting at a GE or Westinghouse with the same amount of background and saying, "We don't know why the heck to do it except that it looks like it's interesting" would very likely have gotten squelched as not being a very worthwhile way to go. There were a lot of ways you could calculate at the time that the economics probably wouldn't make sense or something like that.

Even after we were selling integrated circuits, there were several people arguing that we couldn't possibly build them in economically feasible yields because everybody knew that the yield in an individual transistor was less than 50 percent. If you tried to put eight transistors in a circuit it would be 50 percent to the eighth power, which was uneconomical. I think that kind of argument would have killed the ideas early in a bigger company. It's repeated fairly often, and some of the significant innovations occur in very small companies, little spin-offs without an established bureaucracy to which one has to justify one's work.

Wolff:

TI was a young company.

Moore:

It was young and small by present standards, but it was certainly a lot bigger and lot more mature than Fairchild. However Jack Kilby was very new in the organization and was essentially unhampered during those couple of weeks when he got the original idea.

Wolff:

One thing that intrigues me is this British angle. I have had a little bit of correspondence with Dummer and reading between the lines of a few things he sent me, he feels very strongly that this happened in America because of the kind of incentive system we have here. That the fact that there were young technical guys who could make a lot of money – who could start companies and have stock and this whole entrepreneurial spirit – had a lot to do with it. Without saying it directly he was telling me that being in a big government laboratory in England just stifled everything.

Moore:

<flashmp3>457 - moore - clip 2.mp3</flashmp3>

I believe the latter if not the former. Fairchild was a funny company. There was not an entrepreneur in the group when Fairchild was formed.

Wolff:

You were all entrepreneurs in a way.

Moore:

We were all interested in leaving our jobs at Shockley. We were actually pretty much going to go our various ways and go out looking for other jobs. One of the guys wrote to a friend of his father's who happened to work at Hayden Stone & Co. in New York. He said, "Hey, do you think there is any company that would like to hire several of us? We like to work together and it's a nice congenial group." The guy came running out there and said, "What you ought to do is set up your company. I'll find you support." Such an idea never even crossed our mind. They tracked down [Sherman] Fairchild as someone to support our work in the direction that we thought would be interesting. The first thing we set out to do was hire our own boss. None of us had any business experience and we had no idea of how to organize a company, we started interviewing people to become the general manager of Fairchild. We hired Ed Baldwin of Hughes.

Wolff:

Who was the fellow who wrote to his father's friend?

Moore:

Gene [Eugene] Kleiner. He is now in the venture capital business.

Wolff:

I thought the name was familiar.

Moore:

Kleiner and Perkins. After the Fairchild experience a lot of us became entrepreneurs. None of us were before that.

Wolff:

You as individuals were not entrepreneurs, but the climate was such that the Hayden Stone connection worked and you were able to start a company.

Moore:

Yes. Somebody was willing to put a million and half bucks into a bunch of eggheads, which is sort of what we were at that time.

Wolff:

I think that's the sort of thing Dummer is telling me could not have happened and did not happen in England. He would say that is why he would say they didn't invent it in England. He had the idea, so why wasn't it invented? You say they didn't have the supporting technology. He would also say it's that they didn't have this climate that was conducive to technological innovation.

Moore:

That's true.

Wolff:

That's one of the angles that intrigues me with this whole thing.

Absence of Government Support

Moore:

However the Europeans have a distorted view of the US electronics industry. I was talking at the bureau yesterday with a representative of the German government in the US. He is convinced the reason the U.S. is dominating the semiconductor business today is because of all the government support that is going into it. There is very little government support going in, and what's going in is going principally into the companies that are not very successful – TI being an exception. TI takes some direct government contract support, but this is not case with the rest of the companies. It is, at most, a trivial part of it.

Wolff:

Wasn't it true more so in the '50s and '60s than in the '70s?

Moore:

It may have been true more so at TI. It certainly was not true at Fairchild. Having the government there as a market was important. We could nearly always sell our first expensive devices to some military contractor who needed some new capability, but we studiously avoided any direct government support. The feeling was that if you sign a contract then you commit your people to do something for a year. We wanted more flexibility than that. If we wanted to change our direction, we wanted to be able to do it in a hurry.

Wolff:

I don't think you would have the whole revolution you have today if you hadn't had that military market at the beginning.

Moore:

In the beginning that was useful. To be able to sell your first devices at $150 apiece was important. Of course subsequently the military market has become almost trivial, 10-15 percent of the total now. It is often considered to be a pain in the neck to do business with the government other than for rare opportunities.

Further Reading

Moore, Gordon. "The Accidental Entrepreneur," Engineering & Science, Summer 1994, p. 23-30.

Oral History of Gordon Moore by Craig Addison, Computer History Museum, 25 January 2008.

Oral History of Panel on the Development and Promotion of Fairchild Micrologic Integrated Circuits moderated by David Laws, Computer History Museum, 6 October 2007.

Wolfe, Tom. "The Tinkerings of Robert Noyce: How the Sun Rose on the Silicon Valley," Esquire Magazine, December 1983, p. 346-74.