Lowood: From Rheem/Raytheon, how did you then end up at Fairchild in their research and development laboratory?

Deal: I mentioned at the time, or after the time that Raytheon took over Rheem, that a number of the Raytheon Semiconductor people moved to California from Massachusetts. For about a year, there was sort of a mixing of the two groups of people, and they didn't really mix too well, because the Raytheon people were more of the old school, and thought that they knew all the answers, whereas the Rheem people were the new type of semiconductor people. This all led up to the fact that many, in fact most, of the people from both sides left Rheem/Raytheon, then, in 1963. This was one of the reasons I was thinking of leaving. One of the department managers, Charlie Bittmann, came to Fairchild, earlier, at the end of '62. I finally felt that I needed to make a change. I contacted him. He talked to Gordon Moore and to C.T. Sah, who was the manager of the physics section at the time. Unknown to me, at that time, they were just starting to think about getting a team together to work on the MOS technology. My background in oxidation, and especially the more recent work in the oxidation of silicon, gave me a good advantage in the situation, so I was one of the first of the team to be hired, in early 1963. I also was permitted to bring along my assistant. Her name was Maija Sklar, which caused some problems at Fairchild, because Fairchild, up to that point, did not have any women engineers or scientists. They weren't even technicians, they were hourlies. I had to get approval from Bob Noyce, as I recall, to have her hired as an engineer. She was the first female engineer at Fairchild.

Lowood: You mean it was that explicit?

Deal: There was no written rule, but there was certainly a rule. No women engineers. That was, sort of, the trend in the industry at that time, too.

Lowood: So it was very unusual?

Deal: Very few women at that time went into this type of work on a professional basis. Most of the women we had hired were either hired as secretaries, or as hourly, what we called assemblers in those days. She moved out here with her husband earlier than that, I think. In about 1962 I hired her at Raytheon. She came over here to Fairchild and spent at least fifteen years with me, and only retired from Fairchild two years ago.

Lowood: This group that you mentioned, that was being put together, was it pretty much something that Sah was putting together?

Deal: This was a group that was to be in Sah's area.

[BREAK IN TAPE]

Lowood: This was a group that was formed in Sah's physics division of the Fairchild lab?

Deal: That's right. I was the first one to be employed. I think I came in early March. Another month or so later, Andy Grove appeared, and then Ed Snow, I think, in June. The interesting thing is that we really did not realize at that point that we were going to be working together as a team and help develop MOS technology. Andy, who was more a theoretician in those days, even though he was a chemical engineer from Berkeley, provided the theory. Ed Snow had been working on his thesis at the University of Utah on glasses, which were similar to oxides of course, especially diffusion of ions in glasses. And then I had been working on the thermal oxidation of silicon from the experimental point of view. C.T. Sah was also a theoretician and device person. He was the manager and headed up the team for a while. However, at the end of 1963, he gave notice, I think for the second time, that he was going to go back and teach at the University of Illinois. He had been there once before, a couple years earlier, on a temporary basis. So for a year or two years, it turns out, C.T. Sah was head of the department even though he was located in Illinois, but for all intents and purposes Gordon Moore directed the project. So we would have weekly meetings with Gordon, describing the happenings during the previous week, the results and come up with suggestions as to where we might go from there.

Lowood: This was the department that was known as basic physics, then, in '63?

Deal: That was known as just physics. Earlier there had been a basic physics department as well as an applied physics. Somewhere along the line the basic physics department was dropped and they were combined, with the people that remained in just the physics department. The department that I had applied to earlier, in 1959, was the chemistry department, and it still existed and included both chemistry and materials science, such things as crystal growing and epitaxy and a lot of things like that were carried out in the chemistry department. Worden Waring, who had been the person I interviewed with, was still head of that department, although he soon left, and that department was then headed by Harry Sello. Harry Sello was an earlier Fairchild employee who had gone over to Italy to help set up a Fairchild facility that was owned jointly with SGS. I think Harry returned here in about 1964.
Lowood: Maybe, it might be good at this point to review the departments--at that time physics and chemistry were there, you mentioned materials science was in chemistry?

Deal: When I arrived, or earlier?

Lowood: When you arrived.

Deal: I think by that time the materials group was a part of the chemistry department. Later, then, when I had taken over the department later, it was actually called "Materials and Processes." It was changed from chemistry, even though the content was essentially the same.

Lowood: There certainly would have been a separate group for device development?

Deal: There was a device development department, a very strong department. There was a transducer department, and various support activities. There was, I believe, if not at that time but certainly shortly after there was a department that was called "Digital integrated electronics," which was sort of a product group, helping to develop other types of devices and included computer science as well. It developed into a computer science department.

Lowood: In other words, it was a very large organization for research and development in the industrial sector?

Deal: Absolutely.

Lowood: Do you know why Fairchild went in that direction, that is, to have such a large R & D?

Deal: It's hard to know why the organization was set up with such a large R & D, which was somewhat isolated from the rest of the company. I think they were trying to pattern it after Bell Laboratories or IBM or RCA. I think it was obvious, as time went on, that this separation or isolation from the product manufacturing areas was a great disadvantage to Fairchild. It allowed the people in some of the manufacturing areas to split off, take with them technology that had been transferred from R & D, and then no one else was there to run the production. We may get into that later, but there was a lot of disadvantage of having such a large organization. From the researcher's standpoint, it was ideal. I don't think there has been, and probably never will be, in the semiconductor industry, that type of R & D effort. The closest might be, perhaps, TI. But I think TI is organized so that it is more efficient. The TI research laboratory is right in the middle of their production facility, whereas we were in Palo Alto and the rest of Fairchild was in Mountain View, at that time.

Lowood: You mentioned one thing, I just would like to go one step deeper. You mentioned that there were some similarities, maybe structurally, in the way this was set up, compared to Bell Labs, or IBM. Why do you suppose the people in cHRGe of this company wanted to emulate that model? What was the predilection for it?

Deal: I guess it was because they were interested in having a prestigious organization, and wanted to be big time, is the only thing I can think of. What is very interesting, however, is that when people like Gordon Moore and Bob Noyce formed Intel, they did not organize any sort of an R & D, and even to this day, they do not have a separate R & D laboratory, as we did then.

Lowood: And there hasn't been--well, correct me if I'm wrong, but as far as I know, there hasn't been anything in this region's semiconductor industry that's been like it.

Deal: No. There have been a few companies, like Signetics, who are now owned by Philips and have a separate Philips research laboratory. That's probably the closest. But it's much smaller in comparison.

Lowood: A slightly different question concerning your arrival here at Fairchild. You described briefly the atmosphere at Rheem/Raytheon. What was the atmosphere like at Fairchild, when you got here?

Deal: When I left Rheem/Raytheon, everything was very much up in the air. People were leaving. I couldn't even find anyone to hand in my resignation to, as it turned out. Once I got to Fairchild, I immediately encountered a stable research organization. That allowed many fruitful discussions, and it was quite a different environment, much more like a university. In fact, the Fairchild research center has been referred to as the University of Silicon Valley, over the years, and it certainly was like that in those days.

Lowood: You certainly had the impression things were on the way up for the company, at least for the research and development lab here. I'm also struck that you mentioned that it seemed so stable. That was, four or five years old at that time, right?

Deal: Well, the actual research lab and the large organization only started in 1962. So even though there had been a research group down on Charleston Road before that, it was nothing like the laboratory here. So it was fairly new, about a year old, when I got here. But things were really going. Gordon Moore controlled things very tightly. I don't think he even used a budget other than a budget of the number of people. That's one of the ways that research people have of budgeting, is by headcount. It's very effective, it turns out. I was very happy to get into this type of organization. It's what I had really thought that industrial research should be. I have found, later, of course, that that's the exception, often, rather than the rule. But for a period of five years, a very enjoyable place to work. Not to say that it hasn't been since then, but it was probably at the highest level during that time.

Lowood: Have any impressions stuck with you, maybe at an anecdotal level, of the kinds of things people in the company, in the laboratory said that they were aiming--this is now in 1963--that they said they were aiming to accomplish over the next five or ten years. The goals of the company and the ambitions of the people here.

Deal: Well, I think certainly the ambitions and the goals of management of the company would be to be the top semiconductor company, and they were well on their way. They were not only the leading technical company at that particular point, as far as semiconductor companies went, but they were certainly making the most money and were very successful. That started to change, however, as some of the groups started to pull away and form other companies.

Lowood: But in '63 that was still a few years ahead.

Deal: I think the earliest company after Rheem, which had, of course, broken off in 1959, was GME, later to be bought by Philco-Ford. I think they started in about 1964, '65, taking with them the newly developed MOS technology.

Lowood: GME stands for . . .?

Deal: General Microelectronics. The head of the device development department, a fellow by the name of Norman, got together with a retired Army colonel by the name of-- I can't remember his name now [Arthur Lowell]--formed GME. And GME, later, was the origin of AMI, American Microsystems Incorporated.

Lowood: When you say that, '64, when you say they took the MOS technology.

Deal: '64 or '65, yes. We had just developed stable MOS technology, which allowed the companies to produce it, and they were the first spin off to take that technology and use it. They were not successful, however, as GME. As I said, they were finally bought out by Philco-Ford and then, after that, closed. But some of the people from GME, in the meanwhile, had split off to form AMI. They are still going. They're not, certainly not one of the top companies, but they are still in existence.

Lowood: Then, in fact, already, in the mid--earlier than the late sixties--already in the mid-sixties, these spin-offs were beginning from Fairchild.

Deal: Yes, that's right. There were several companies, if you look at the genealogy chart, that broke off before National and Intel.

Lowood: As you mentioned, you went into the MOS research group, and you mentioned your collaborators, Grove, Snow, and Sah, that you eventually came together, realized you were working on the same things. Was that group constituted formally at any point within the organization, or did it remain a kind of an ad-hoc thing.

Deal: The three of us, Snow, Grove, and myself, all reported to Sah, so we were a group within the physics department, and we definitely had a charter of developing technology for producing stable MOS. I keep referring to stable MOS because it turns out that the biggest problems with MOS devices that had been developed and invented at Bell, two or three years earlier, were very unstable, and there was some source of instability, apparently, associated with the oxide, which was part of the planar process, and also the gate for the MOS device.

Lowood: OK, we'll get back to the organization of the company, and I'm going to continue with this, with the MOS project, since you've started to talk about it now, and I think it's a good time, probably, to talk about it. Were those problems, then, the ones in 1963 when you arrived at the company, were those the ones that you started to work on, or did you work on the experimental work?

Deal: I had been working, along with one of the researchers that had come from Raytheon back east. His name was Dick Hager. He and several other of the Raytheon people were quite knowledgeable in the surface chemistry and physics of both germanium, and then later silicon. So I really got my background in electrical characteristics and stabilization of oxides through them. Once we got here, it became obvious that the same problems that we'd experienced over there, that is, having the electrical characteristics of an oxide on silicon remain stable, was going to be a problem as far as MOS transistors. About this same time, in fact, one of the first projects that Grove and Snow and I worked on, were to develop the characteristics in a model for MOS structures, metal oxide semiconductor structures. This is where the capacitance-voltage method came into being, and our first successful paper, along with C.T. Sah, had to do with the C-V characteristics and the theory of C-V characteristics of MOS devices.

Lowood: So this is the paper you were referring to now, "Investigation of Thermally Oxidized Silicon Surfaces Using Metal-Oxide-Semiconductor Structures" in Solid State Electronics in 1965.

Deal: In that paper, the theory and background for the use of the C-V method of analysis for investigating electrical properties of thermal oxides on silicon was developed, and it was recognized as a standard paper. In fact, it became the first of four papers that were designated as publication classics by Current Contents Magazine.

Lowood: So that you've written one of the short pieces for that?

Deal: There are four of those.

Lowood: I'm not familiar with the method. Can you describe, in a general way, what the capacitance-voltage method of analysis was and how it was applied.

Deal: Well essentially, a capacitor, which includes a silicon substrate, a thin dielectric, such as silicon oxide, and a metal field plate, which in this case was always aluminum, is prepared. Depending on the preparation conditions of the oxide, how it was formed and so on, will determine what electrical characteristics you have. What happens is that you are measuring the field as a voltage, applied across the capacitor, and you're also measuring, at the same time, the AC capacitance. So you're really measuring the AC capacitance, which is the Y-axis of a plot, against the applied field, which is the X-axis. At some point as you change from, for instance, large negative bias on a P-type semiconductor substrate, to zero bias and then change to positive bias, you invert the surface, and you have a decrease in the oxide capacitance that's measured, because a space-cHRGe region is formed in the silicon substrate, and the effective thickness of the dielectric, now, is both the thickness of the oxide plus the space-cHRGe region. In that case, capacitance drops. The position of that capacitance drop along the voltage axis compared to a corresponding theoretical plot, is a measure of the number of cHRGes in the oxide. The more cHRGes, the more positive cHRGes in the oxide, for instance, the point where the capacitance drops along the voltage axis will move to the left, to more negative values. (Figure 1) If you have, for instance, negative cHRGes in the oxide, the curve will be shifted in the positive direction compared to the theoretical plot.

It turns out it's a very simple method for determining cHRGe formation, the presence of cHRGes in oxides, in oxidized silicon, but it's very quantitative. Not only that, but if there are mobile cHRGes, such as sodium, in the oxide, and this is what we found to be the cause of the instabilities, primarily, you can measure the position of the C-V curve along the voltage axis before and after moving the sodium ions from the field plate interface to the silicon oxide interface using a bias-temperature stress test. You can move them back and forth and watch the C-V curve shift back and forth. And then, as I said, you can quantitatively relate the shift of the C-V curve to the amount of cHRGe in the oxide. This gives you a tool to measure the effect of all sorts of process conditions on cHRGe formations. Stress conditions are normally ±50V/um of oxide, 300o C, and 2 minutes.

As a result of these studies over the years, over about a ten year period, clear up into the early 1970s, we were able to characterize four different types of cHRGes associated with the thermally oxidized silicon, one of these being mobile ions like sodium. Others were fixed cHRGes. Another one was an interface trapped cHRGe, or what used to be called surface states, and the fourth was cHRGes trapped in the bulk of the oxide. There are many process variables that will still affect each of these cHRGes, so a very complete picture has emerged, over these years, of process effects on cHRGes and oxides, and their effect on device characteristics.

Lowood: So this was, in effect, a way of reading from the graphs, the capacitance-voltage graphs, you could read what was going on in this as yet, at that time, not fully understood system, you could isolate various effects.

Deal: A very simple but effective method of characterizing the electrical properties, such as cHRGe properties of oxides, and it allowed us to determine, for instance, that the instabilities that people had noted over these years in thermal oxides and MOS transistors, were actually due to mobile ions, like sodium, lithium, potassium. Very interesting series of events were occurring, in that there were several laboratories working on the same problem. IBM, RCA, Bell Laboratories, Philco and a couple of others, and we had a very high degree of competition among these groups. In certain cases, there would actually be, not animosity but close to it, between or among the various groups. I know RCA was very volatile and Andy Grove tended to be very volatile in his discussions and opinions.

But it turned out that the primary cause was the one of ion migration, alkali ion contamination, which was the one reported and considered to be the likely one from the Fairchild side. An interesting story was associated with this competition that went on. During the time, or just shortly after it was decided and agreed by most groups that sodium and alkalis were the major cause of MOS instabilities, a fellow by the name of Bob Donovan, at the Research Triangle Institute, presented a paper, and in this paper he had a picture of an elephant with a bunch of blind men around it. Now there's a poem, a famous poem called "The Blind Men and the Elephant." The gist of the poem was each of the blind men was looking at a different part of the elephant, and each thought that the elephant was something different. The one that was feeling the trunk thought that the elephant was something like a flexible rope or something. The one feeling the legs thought it was like a tree, and the one feeling the ears thought it was like a palm branch or something. In reality, each, it said, was partly right and each was partly wrong. On this drawing, Donovan showed each of the blind men to be one of the groups that were working on the oxide instability problem. And Andy Grove had heard about this elephant and the blind men, and for several weeks he was very concerned as to where he was positioned on the elephant. I know Bob Donovan from previous association, so I called him and asked him to send the picture of the elephant so we could straighten this out, and when Andy Grove found that he was shown on the head of the elephant and one of his arch-rival Hungarians was shown on the tail, you could hear his shout of glee all over the whole valley area.

Lowood: When you talk about the volatility of this competition, I take it that there was communication of results but disagreement about their interpretation.

Deal: That's exactly right. There would be workshops, IEEE would have workshops. We had several meetings at the Electochemical Society. Other meetings in which each group would present their particular ideas as to what some of the causes of these problems were. Not only the instability problems but some of the relationships of some of the other cHRGes and nature of MOS structures in general. And naturally, it was a new field and there were a lot of disagreements. For the most part, people got along pretty well. There were a few instances where we thought that maybe things could have been handled a little better.

Lowood: Secrecy, though, I take it, wasn't . . .

Deal: It was and it wasn't. Secrecy was important in certain cases, but from the scientific standpoint, we communicated fairly openly. But it wasn't until we had obtained patents on how to control and minimize alkali formation in oxides that we really talked about it. So the patent situation continued to be an important aspect of companies like Fairchild.

[END OF SIDE A, TAPE 1, BEGINNING OF SIDE B, TAPE 1]

Lowood: That is, on this relationship between the scientific communication and the patent situation. I notice, and perhaps this doesn't mean very much, that the papers that came out of your group always referred to the MOS structure and always referred to a physical system, but that there's very little in the way of discussion of devices.

Deal: That's right. We purposely did not talk about devices and device characteristics. I think there is only one paper in that list that talked about reliability studies of MOS devices. It also happened to be, though, that I think the people doing the more, if you want to call it fundamental research, were the type that would publish their papers, whereas the development engineers, the production engineers are not the type who publish their papers as much. So that was another case. But I think they would have been restricted by our patent department.

I should point out one other important oxide research area in which I was involved. That was thermal oxide kinetics, which had been started at Rheem. With the help of Andy Grove, we developed a standard mechanism for silicon thermal oxidation--the linear parabolic model. We published the paper in 1965, and the mechanism is now referred to as the Deal-Grove model. Since 1965, the paper has been cited an average of 30 times a year in the literature. It was used later in the SUPREM program. (SUPREM will be discussed in the third interview.)

Lowood: Were you also working on devices at all, and just not publishing?

Deal: We did not work on devices ourselves. The device development groups that we had, a couple of them, did work on products.

Lowood: What was the interaction, then, between your group working on the structure, and the work of the device development group?

Deal: We worked together. The weekly meetings that I mentioned earlier that Gordon Moore held were attended by device development people as well as ourselves. We talked about problems, and the people developing the products would discuss some of the problems they would have and ask for assistance, and we would go back and try and work out some of the details, find out what was going on.

Lowood: How would that work, let's say from the point of view, now, of the device development group? Were they working on ideas concerning MOS devices right beginning at the same time as your work on the structures or was there a delay?

Deal: There was a slight delay, I would say. We had to work out some of the details of just being able to characterize, produce, get the right kinds of oxides, thicknesses of oxides and structures. But fairly soon after we started and worked out the C-V method of analysis and so on we had a device development group working on the problem.

Lowood: Was the sole means for communicating the results of the research side to the device development side these weekly seminars? Of course, they read your papers, obviously, but were there other means?

Deal: Well, we were in the same building, and as long as the development group was in the same building as the research, then you'd have communication directly. The problem starts, and it started with Fairchild, when the development groups transfer their products to a production group, which was in a different building. Two things happened. One thing, in that case, is that you do have less communication, because it's more difficult to get back and forth. The second thing that happens it that these are different people, typically, in the production groups, and they decide that they want to change the process and that they can produce it better than the development group had produced it. And this has been a problem with semiconductor companies from the start and it is today. How do you transfer a product from one group to the next? Some people will move people, will move key people of the group into the production area and they stay there for a year or two. That's probably the best way to do it, because they are the ones who worked on the development end of it, and are not going to be changing what they've already done. The difficulty is, in this country especially, is that people in R & D, research and development, have not wanted to go to production.

Lowood: In the case of the MOS developments here at Fairchild, then, you're saying that the communication, because it was fairly informal, was good between research and development?

Deal: That's right.

Lowood: They were fairly in step and results were communicated fairly easily, but that once it got past the development point, that was where communication began to become a little bit more tenuous.

Deal: Not only that, but once a process had been transferred to production, it would be quite common for that whole group to leave Fairchild and start a new company. That happened several times.

Lowood: I'm a little curious about one thing, this notion that somebody in production would think they knew a better way to do it. Was that based on a feeling that superior knowledge, or was that based on certain things maybe not working conveniently?

Deal: It's based on human nature, as far as I can see. Everyone thinks that they can do it better and that they have a certain idea that they think would work better than someone over in R & D.

Lowood: What kind of mechanisms, solutions, whatever, were attempted in the company to avoid this problem?

Deal: Well, the one was to transfer people. So we did transfer people, as long as the production was in Mountain View, which is still not a long distance away from Palo Alto, we had people going back and forth. Even after production went to other locations in the company, we've had people coming back from the production group to work in R & D on new product, even within the past five years. They spent six months to a year and then went back with it. This has also been fairly successful.

Lowood: I think there was one point in your resume that you were in cHRGe of a production group?

Deal: I was in cHRGe of MOS engineering. It turns out that that was a fairly sudden thing, because the man that we had in cHRGe, Bob Seeds, who had been the department manager in R & D, was killed suddenly, and they needed someone to take over that job, since it was one of the crucial times. It seemed like all of the times were crucial for Fairchild MOS organizations, but I was asked to go over. I spent a year heading up the engineering for the MOS department, as it was called, or division.

Lowood:And that was in the mid . . .

Deal: . . .seventies.

Lowood: I notice here, it says also "Responsible for liaison between R & D, CCD, CMOS, and Wappinger Falls MOS technologies.

Deal: That's right. That engineering manager for the MOS division had that responsibility, to help to coordinate the work going on in the various production groups, as well as in R & D. Another attempt to improve communications.

Lowood: This was, of course, ten years into the game, so this was an ongoing problem through all these years. One more question on this. Who would have been, as an individual, responsible overall within the organization for coordinating communication between groups, between areas like research and production?

Deal: Well, that would have been the director of R & D, initially Gordon Moore. He would have the responsibility for seeing that the different groups in R & D communicated with each other, worked together. He also would have had the responsibility for seeing that there was communication with various production groups at Fairchild and R & D. And I guess, then, the next person responsible would have been Bob Noyce, who was general manager. Gordon Moore reported to him.

Lowood: And the head of production during those years?

Deal: One of those people was Charlie Sporck. We had several production general managers during those earlier years. As a group would leave, another manager would have to be appointed.

Lowood: The only other large group I can think of, that we haven't talked about that might have a role in this would have been marketing. Sales, I guess, would be a better term in these years. That would have been Jerry Sanders during these years?

Deal: He started out as a salesman, and I guess he ended up as general manager of sales, marketing. R & D, especially people in my position, had very little to do with that group. There was supposed to be some sort of an organization, a product marketing and coordinating group that would work with R & D and production. How successful that was I'm not really sure.

Lowood: Here's a very general question, but I suppose what we've been talking about leads up to this question, which would be who would then decide which products to emphasize, which products to make a larger research and development effort in?

Deal: This is always a combination of input from the researchers themselves, from the director of R & D, from product marketing groups, the production area, applications people. That's one of the things that I think, one of the places that companies fall down on the most, is not having a good coordinating group to decide these things. From the market standpoint, from the technology standpoint, from the business standpoint.

Lowood: So that was the case at Fairchild?

Deal: There was an organization set up to do that, but I really can't recall how successful it was. It probably wasn't as successful as it should have been, obviously, if you look back at the record.

Lowood: It must have been on occasion the case, I'm thinking here of interviews that have been done with people at, say, Hewlett-Packard or Varian, there must have been some projects that were considered central to the organization, and that the top people in the organization felt were important, and therefore, beyond discussion, and I would guess the whole MOS and the planar transistor area, would have been considered the core area of the company during these years?

Deal: Well certainly, yes. The planar process, by this time, even though it had only been invented two or three years old, was the whole basis for producing semiconductors and integrated circuits.

Lowood: Were there times when directions within R & D changed because of a directive from high in the organization?

Deal: Yes. I can think of later, when I believe it was Wilf Corrigan decided to get into consumer products, which the company had never gotten into. We were supporting the consumer products division and the optoelectronics division in the manufacture of watches and TV games and all sorts of things. A lot of the work that was done in R & D was supporting that effort. This was directed by the president of the company.

Lowood: That was in, when, the late seventies, right, mid to late seventies?

Deal: Yes, right.

Lowood: That wasn't the style that Moore and Noyce adopted?

Deal: No. In the earlier days, in the sixties, it was pretty straightforward, the types of products were pretty well established. I think they were going fairly much with the general business trends. They certainly did not get out of the mainstream. They did, however, and this is an interesting aspect of that, get into gallium arsenide products and were one of the leading gallium arsenide producers, as far as devices, both microwave and gallium arsenide FETs. This was in the late sixties and early seventies.

Lowood: You're saying that was something where Moore in particular made that push?

Deal: I'm not sure whether it was Moore or Noyce. I'm not sure where that came from, but that was certainly set up at the top level of the company.

Lowood: Was there any particular reason?

Deal: I don't know what the reason for getting into it. They really get into it almost from the time that I started. But a business decision was made in the late sixties, early seventies to get out of it, because they couldn't afford to stay in so many areas of business.

Lowood: I want to return for two or three more questions to MOS development. One was a question about the kinds of pressures there must have been to bring something to market. I guess that occurred in '64 with the PLANAR-II. Is that correct?

Deal: In '64, which was the result of our efforts in making stable MOS transistors. The PLANAR-II was a term that was used to denote stable MOS. And yes, there were certainly, but right at that point was when GME took the process and went somewhere and took some of the best people. And then it wasn't very long until some of the other spin-offs started as well.

Lowood: So it was called the PLANAR-II to indicate that the problems, the earlier problems with stability and so forth had been eradicated.

Deal: Right. PLANAR-II really had nothing to do with planar, other than the fact that it was used in a planar process. But because Fairchild had been so successful with planar, they wanted to use the planar designation, and indeed, they even used it later as isoplanar, for the first oxide-isolated bipolar integrated circuits. It was sort of a marketing, advertising type of thing, really.

Lowood: Did you have anything to do with, did research, your end of research have anything to do with the last-minute things in bringing the product to market, the testing and last-minute developing, the tweaking and all of that sort of thing?

Deal: No. Once it went from R & D over to production, the engineering organizations in the production groups did most of the reliability studies as well as the final determination of what specs were to be considered, and so on.

Lowood: Now you've already discussed some of the problems with the early MOS structures or transistors. You talked about the voltage instability and the kinds of techniques you brought to bear on that. Were there other innovations that secured Fairchild's place in the development of MOS technology, as being a leading company, that we haven't mentioned yet?

Deal: I don't know if I mentioned the fact that while we knew what the cause of instabilities was, and that was a basic advance, the next thing was to develop the processes and the understanding to make the devices stable. There is a patent which I am one of the authors of which goes through some of those steps, like using a clean metallization system, that is using electron-beam metallization, rather than filament metallization, which had been used previously. There were other steps along the way in which you would minimize the sodium incorporation into the process, and then finally, there were so-called gettering steps, that is, steps where some sort of a glass, in this case a phosphorus oxide, be used to actually suck the impurity ions out of the oxide after they got in there. They got in there after all of your attempts to keep them out. So all of these details of processing that were developed, and these are some of the things that we did not talk about so much. We'd talk about that basic physics of the instability, but we wouldn't talk about the cures of the instability as much. For that reason, in many of the papers that are reported, you will not really see papers dealing with some of the details of solving the problems. That's a different sort of thing, what you can report and what you can't report.

Lowood: This gets more into the trade secrets side of it.

Deal: That's right.

Lowood: The development of these process technologies, in order to remove the various impurities and the other things, was that something that your background prepared you for better compared to some of the people with a straight physics background, for example.

Deal: Well, it was more chemistry rather than physics or electrical engineering, and it brings another interesting note to the discussion is that the people that made up these groups in the early sixties, and still do as far as that goes, come from varying backgrounds--physics, chemistry, electrical engineering, materials science, even geology, mathematics, and so on, and they each bring something to help solve problems, from their own particular background. On the other hand, you will find, in many companies, all these people mixed together, and you could not tell, by looking at the job they're doing, what particular background they really came from. Obviously, a designer would have been an electrical engineer, but with someone working on one aspect of a processing problem and someone else on another one, it's very difficult to tell what background they came from. My own feeling is that chemists have probably been underused in the industry, and electrical engineers have been overused. In a lot of cases, electrical engineers have been assigned to processing problems where they really didn't have the background. A chemist could have done the job better. Unfortunately, universities have tended to have all of their solid state work in the electrical engineering departments, and to some extent materials science departments, and not in the chemistry departments. There's not more than one or two chemistry departments in the country that really specialize in solid state. It's hard for me, as a chemist, to accept, but that's what has happened.

Lowood: In these process areas, it certainly seems fairly obvious that a complete understanding of what's happening as you introduce additional elements into the system would require at least some background in chemistry, because you could very easily be introducing something as you're taking something else out of the system.

Deal: A good example of that is, as I said, filament systems, filament evaporators, where aluminum is evaporated from a heated filament, caused instabilities. We didn't know why and we went to a system which involved electron-beam evaporation, which tended to distill the impurities out of the aluminum and deposit more pure aluminum. Later, we found that the filaments, which were generally tungsten, were made by a process that employed potassium, and potassium is almost as bad as sodium in instabilities. Once that was known, manufacturers were able to come up with a different process to have higher purity tungsten filaments, but we never really got back to them, because electron beams just took over.

Lowood: That raises a general question, that I suppose a little bit has been written about, but I wonder about it. In these kinds of innovations, the innovations that result in the equipment that comes into a large operation like Fairchild. Generally that equipment is produced by another company?

Deal: In the early days, it was produced by the semiconductor companies, and much of the equipment that we used was produced at Fairchild. In several cases, those people then left Fairchild, with the support of Fairchild, to form separate companies to produce the equipment, since again, in those days, Fairchild did not want to get into other types of business. A good example of that are Kasper aligners, I think were developed originally at Fairchild, and the people who developed those formed a company called Kasper.

These are aligners for aligning the wafers during photolithographic operations. Some of the early furnaces--when I came here all the furnaces were built by Fairchild at Fairchild. Fairchild supported a couple of small spin-offs to form separate companies to produce furnaces. The aluminum evaporators. I mentioned we went from filament evaporators to electron-beam evaporators. It turns out while electron-beam evaporators are very clean, they also produced radiation, which in later types of devices were not satisfactory. So we went to a third type of evaporator, called a flash evaporator. The first one was developed here, and then we transferred the rights to a local company to produce those flash evaporators.

Lowood: So it really wasn't all that different from the development of devices. These equipment developments occurred . . .

Deal: . . . along with, yes. It was required to develop new types of equipment. We had to do that too, but we did not stay in the production. The so-called Vapox process, for low-temperature deposited oxides, the equipment was developed at Fairchild, and then later, the rights were sold to, I think, Pacific Western, and picked up by Applied Materials, I believe.

Lowood: The question I was going to ask, and I guess it doesn't really apply, then, to this early period, but maybe it does a little bit later, is if a company that produced that kind of equipment changed things, for one reason or another, and I suppose it could maybe just be economics, was that usually communicated?

Deal: No. And another interesting thing, along with the companies, most of the equipment companies are very small and do not have a lot of money for R & D. They depend on the larger semiconductor companies to do the R & D, even after the equipment has been sold to them. You have a first generation of CVD reactors, a company like Fairchild or Intel or National, any of them, would buy those reactors, find out problems, they would report back to the company, and the next generation would incorporate the improvements. And in many cases, there would actually be cooperative programs going on between the equipment company and the semiconductor company to help improve and develop new types of equipment. It's very common, even today.

Lowood: And that's really the mechanism that, I guess, creates Silicon Valley in a way, is that all of these companies that service various needs of the larger companies begin to be created.

Deal: In the case of Japanese companies, most of the equipment companies are owned by the same conglomerates that own the semiconductor companies plus the chemical companies that supply the chemicals and so on. It's all one big huge company, and that's the way they do it. Even tighter control, I guess, on what's being done.

Lowood: In this connection, I had one question that is a little bit off the track here, which had to do with the development of this technology of wafer spinners for the resist coating. I had just been reading about that, this one author talked about how spinning platters have been developed in house here at Fairchild, around 1962, but said any record of these developments had pretty much disappeared.

Deal: I was not familiar with that particular development, like some of the others, but I'm sure it did happen. But absolutely no records, you're right. Only in the minds of the people that were involved.

Lowood: Can you recall, for this record, any important undocumented technologies that were developed in your work?

Deal: Well, as I said, low temperature CVD of oxides, Vapox, was certainly one. I mentioned the aligner. Kasper was one of the early companies to produce contact aligners for aligning masks to produce semiconductors. Furnaces, I mentioned, were first built at Fairchild, and then the designs were either sold or given to people outside. There may have been a couple spin-offs of furnace companies. I think there were. Thermco, I think, was an example of that. The evaporation, the flash evaporator was one. All of these. There'll be no documentation that I know of, unless, perhaps, in the progress reports that we have, something might be indicated there. It would be interesting to look for that.

Lowood: Would any of these have been developed in your group, sort of, under your eye?

Deal: Vapox was, for instance. That was done in the chemistry and materials and process department. Mike Barry was the person that was responsible for that. As I said, that was transferred then--the rights to produce that were transferred to another company.

Lowood: Can you talk a little bit how it was developed, how it came about, just so that we have an idea of how this process worked a little bit?

Deal: In certain cases, you want a vapor-deposited oxide, where the oxide is produced in the vapor phase and deposited over the device or circuit. Unlike thermal oxidation, which uses part of the silicon to convert to an oxide by a reaction with oxygen. The process that we were using when I came here was called TEOS, which was tetraethylorthosilicate.

[BREAK IN TAPE]

Deal: A process which involved the breakdown of a organic oxygen-containing compound to form SiO2. It also contained silicon, obviously. This was done at about 750 degrees. The problem with this process is 750 degrees is way over the melting point of aluminum, and we desired to have something that would go down under 450 degrees, be deposited at less than 450 degrees. Somehow, the researcher who was working on this problem, working for a solution, had heard or know about that you could react silane, SiH4, with oxygen, to produce SiO2. Again, a vapor-deposited process, but it would be deposited around 400 degrees. That was the driving force, was to be able to have a deposited oxide that would go over an integrated circuit, having metallized aluminum as the interconnect material. There's some early papers in the early sixties on that process, developed here at Fairchild. Now it's quite a standard process. The only thing that's happened to it is that people now have added plasma, activated process with plasma, which produces a more even, uniform oxide.

Lowood: Then as you mentioned, that would be an example of something
that worked its way out of the company.

Deal: Out of the company and into the industry.

Lowood: Let's go back now to the organization of the company. We've talked about the MOS project now. I'm referring to this Fairchild organization chart from 1968, and you've given me this other document from '62?

Deal: '62. That was the opening of the research laboratory. Right in the middle is a description of the various departments that were in place at that point. Some of them were the same as are indicated later, in 1968, but some had disappeared. I think this is the case with any organization, especially it was the case with Fairchild's research laboratory, in that as people came and went and as times changed, why, maybe one department would be eliminated or two would be combined, depending on the requirements of the organization at that point.

Lowood: That was something that happened quite a bit in the company, I imagine because it was a very dynamic company at that time that there were lots of changes.

Deal: Since I first joined the semiconductor industry to now, the organizations are always changing, both in R & D as well as throughout the companies.

[END OF SIDE B, TAPE 1/BEGINNING OF SIDE A, TAPE 2]

Lowood: Could you describe the groups and mention some of the lead people in each of the groups, in the sixties.

Deal: Well, of course, Gordon Moore was the director of R & D at the start, when this building was constructed in 1962, or this part of the building was constructed in 1962. He was one of the founders, as was Dr. Victor Grinich, who was associate director at that time. There were several departments. One was transducer department, headed by Jack Kabell, that later transformed into some sort of a sensor or a light-emitting diode department, the forerunner of CCDs. The chemistry section was headed by Dr. Waring. This department had not only physics and chemistry processes and so on, glasses, but also then absorbed materials work very shortly after 1962. That is, research into crystal growing and the like. Another department was a microwave physics section headed by Dr. Irv Solt. Microwave physics got involved fairly early in gallium arsenide devices, which Fairchild was quite well-known for, in those days.

Lowood: Excuse me. These would be power devices or what?

Deal: No. Microwave are very high-speed devices. And that's why they went to gallium arsenide fairly soon.

Lowood: So these would be microwave components?

Deal: Right. Watkins-Johnson, for instance, is a company that makes microwave devices even today. Basic physics section was something that had a short life, as I recall, headed by Dr. Herb Leifer, but I think basic physics was a little too basic for a company like Fairchild, didn't last very long.

Lowood: That would mean solid-state physics, something like that?

Deal: Even more basic. It was not clear to me just exactly, they were supposed to be working on things way out, way out in the future, and even Fairchild just couldn't support that sort of thing. Certainly solid-state was involved in that, yes, but there was a solid-state physics section which was more application, more applied type of work, headed by Dr. Tom Sah, and that was the group that I was hired into. There was also a high-speed memory engineering group, headed by Harley Perkins. That group was the forerunner of other development groups of specific types of devices like memory, we had memory development all through the sixties. Those were primarily the main departments when the lab was set up here in 1962. And then, as I said, the chemistry section absorbed materials work, or took on materials work, and was changed to materials and process. The basic physics was eliminated and most of the work, then, was done in the physics department. The transducer section was changed over more to imaging type devices. The high-speed memory engineering section changed names several times, and was involved in various types of things. The device development department continued, pretty much, over the years, as a device development, in which the main line, or pilot line, for the research center was in that group, and they produced devices for all the other departments. Each department would not produce devices themselves if they needed them. They would go through the pilot line and device development. The person that headed up the pilot line, was Don McCall, who just retired this last week from Fairchild. He later went into facilities and did an outstanding job over the years in helping to install new equipment in this facility.

Lowood: Besides Gordon Moore, obviously, who were, over a period of years, who were really the key managers within this organization?

Deal: You mean as far as the director himself?

Lowood: As far as both administratively, as far as sitting in a high place in the organization, but also who were recognized to be the real leaders.

Deal: After Gordon Moore, Jack Kabell, who had been a department head, became director. Harry Sello, who was in cHRGe of the materials and process department then was made manager of planning, and then he later got into international relations-type work, in helping to establish trade agreements with Poland, Romania, China and other places. He had left the company about five years ago and went into consulting. A man by the name of Rex Rice was instrumental in a lot of the development of some of the digital devices and systems application, for related products. A number of the Intel people, like Ron Whittier and Des Fitzgerald, of course Andy Grove, Will Kauffman, Federico Faggin, Dov Frohman, who's now known as Dov Frohman-Bentchkowsky. All of these people helped develop a lot of the technology. Pierre Lamond took over from Phil Ferguson as head of device development and later went to National, and I'm not exactly sure what he's doing now. Gil Amelio, who is president of Rockwell Semiconductor group, headed up the cHRGe coupled device group here. That group is still going. In fact, the only group still operating in this building right now. Will Steffe was later device development manager and then manager of Fairchild's high rel line in South Portland. Jim Early was hired by Dr. Les Hogan, when Hogan came from Motorola in 1969. Jim Early was director of the laboratory for fifteen years or so. I'm sure there are other who I have not mentioned, but there's a whole series of people, of course, that have been leaders in the industry, either first at Fairchild and then went on to other places.

Lowood: So was it then Moore, and then Kabell and Early were the three directors?

Deal: Those were essentially the first three directors of the lab. In very later years, Tony Ley. When Schlumberger bought out Fairchild, Tony Ley was assigned to that job by Schlumberger. So in the last five or six years, Tony Ley was vice-president and director of the R & D lab.

Lowood: Can you contrast the styles of Moore, Kabell, and Early? Did they have, they must have each had a different imprint on the organization.

Deal: They certainly did.

[BREAK IN TAPE]

Deal: In comparing the different management styles of the directors we've had over the years, it's an interesting contrast. Gordon Moore, of course, is one of the early leaders in the industry, a founder of Fairchild, and is much sought after today to give opinions and views on where we've been and where we're going, really one of the senior statesmen of the industry. Jack Kabell was an interim manager and director, and I guess we can't say too much, because he didn't have much time to prove himself there. He left the company shortly after the year that he was director and went into consulting. He's now in San Diego. When Les Hogan took over the presidency of Fairchild Camera and moved the headquarters from New York to California, he also brought in Jim Early from Bell Laboratories, whom he had worked with at Bell earlier. Of course we've talked about these people, many people who have been at Bell have been leaders in the semiconductor industry, except for the founders of Fairchild. Jim Early was very brilliant. He had developed some of the early theories on semiconductor device applications. There's an effect in bipolar transistors known as the Early effect. It's quite well known today. He brought some of the background with him from Bell, more recent background and ideas and so on. He was director for about fifteen years. He was certainly a different type than Gordon Moore, his main predecessor, and certainly a different type than Tony Ley, who came after Jim Early. When Fairchild was bought by Schlumberger in 1979, then they brought in shortly after, Tony Ley who had been director of a solid-state laboratory in Paris, although Tony did not have any direct semiconductor experience, he certainly learned about it quite rapidly, and he did have a lot of circuit knowledge, I think, systems in general, which helped complement the device and technology experience the people here had.

Lowood: The basic development of the organization, though, wasn't changed by any particular philosophy that any particular director brought in? I mean, were there major differences in the philosophy of research and development at Fairchild?

Deal: Not really. If there were, as we changed over the years, it had nothing to do with the director. It had more to do with business conditions and the need to either cut down or expand as we did at various times along the way.

Lowood: So within the limits of developments in the industry, the early decision, I guess, to have a large research and development operation, to, as you say, quasi-model it after Bell Labs, pretty much was maintained through all the dips and rises?

Deal: Well, there's one big exception. Shortly after Jim Early came here in 1969, maybe about a year later, the business really took a turn for the worse, and the president of Fairchild at that point, who was then Wilf Corrigan, who had taken over from Les Hogan, determined that we could not afford such a large R & D. In a period of about two weeks, we cut R & D from about five hundred people to fifty. I myself had the unpleasant task of laying off a hundred people. However, of those hundred, and of the other three hundred fifty people that were really technically out of a job, we found jobs for all of them in other parts of Fairchild, and a number of those people then later came back to R & D as we built up again, which happened. So we stayed at fifty for a few months or so, but then we started building up again. We brought in some other groups into this building, production groups, to help support the building's operation. But they gradually, then, were transferred out again, and R & D people took their places again.

Lowood: And the way you presented that, it wasn't that Corrigan was hostile to R & D, it was that something had to give in the company.

Deal: Well, everyone didn't agree with Corrigan's philosophy in those years. And they didn't in later years, either, in certain cases, when we got involved in the consumer products. However, personally, I always found Wilf Corrigan to be more supportive of some of the programs that I was involved in than some of the presidents you might have expected to be more sympathetic. Very interesting, you can't tell.

Lowood: From your position, during these years now, in the sixties, you were moving up in the organization, and also to a certain extent, correct me here if I'm wrong, you were moving, a little bit into different aspects of the company. This one year that, well, this was in the mid-seventies, that you were involved in production.

Deal: I was involved in production. That was one thing I should point out, that Fairchild sort of expected their managers to do, both in operations parts of the company and the R & D facility, was to get some experience in other parts of the company. Which was one of the reasons, not the only reason I suppose, but one of the reasons why I did move over there and it didn't disturb me that much.

Lowood: I noticed also that you went from managing materials and processes, which makes a lot of sense, in the late sixties and very early seventies, to becoming manager for device technology for most of the seventies.

Deal: It was the same department, though, that was a combination of the physics department and the material and process department and ended up to be device technology. And we seem to change names periodically for various reasons, I don't know why.

Lowood: That sounds very different from materials and processes.

Deal: Well some of the people were the same, but we did combine the physics and the materials and processes.

Lowood: Was there device development in that group?

Deal: And at one point, even device development was combined. I had a very large organization. It was maybe because things were uncertain and it wasn't clear. They didn't want to bring anyone else to head another department at that point, so I took both of them over.

Lowood: As long as I'm asking this question, there are a few dividing lines I wondered about. We've talked about cross-department and cross-division communication within the company, and your feeling is that, at least within research and development, communication was informal and that it generally functioned fairly well. There were a couple of dividing lines, nonetheless, that I wondered about. One was, and now I'm thinking of the organization as it was in the sixties, the dividing line between physics and then the materials and processes, which was more on the chemistry side. What determined whether one person was in one group or the other? What difference did it make?

Deal: Well, it didn't. The interesting thing, though, as I recall, there was some competition between these groups. I remember in trying to determine how to make stable oxides for MOS devices, there was a group working in the chemistry department on that same problem, and there was some competition, and sometimes it got a little warm in some of the discussions people had. Of course, the thing that solved that whole problem was when I moved from the physics department and took over the materials and process department, and I think most of that sort of stopped.

Lowood: What stopped? The disagreement, competition?

Deal: Any disagreement there might have been. And there were a couple of people that maybe were more inclined to encourage that, and it turned out both of them left, one from each of the two departments.

Lowood: Maybe I don't need to ask about a lot of these distinctions. It seems that what you're saying is that the boundaries were not really that strict.

Deal: No. We had people moving back and forth between departments.

Lowood: And you've given some examples of when there was cooperation as well. One thing I was curious about was, where in the organization technologies like mask reduction and mask graphics and mask alignment, those kinds of things having to do with integrated circuits and all?

Deal: That originally started in R & D and was part of the materials and process department for a time. They reported to me at one point. Then it was determined, and I don't know exactly how that determination was made, thinking back, that the mask department, mask-making department would be moved to Mountain View, into a separate building. I think part of it was undoubtedly due to the fact that we didn't have the facilities here, we were so crowded. And for the new types of step and repeat cameras that we were using in those days, you needed a more stable facility. Not only we didn't have the space, but this facility, on the first floor, was not very vibration-free, and there wasn't any place in the basement to put it. So we transferred those people there, transferred the organization. It has always still reported in to R & D, administratively. Maybe for a while it didn't, but typically, the majority of the time, it has. Up to the time that Fairchild was sold to National, the mask-making shop reported to R & D, to the director.

Lowood: To the director of R & D directly?

Deal: That's right.

Lowood: Some of this involves things that happened before you came here, I suppose, and in a way maybe this is a silly question, but seeing Noyce's development of the first integrated circuit. That occurred at the top of the organization, as it was, I suppose, at that time.

Deal: The organization was pretty small at that point. I think everyone was involved in that. And I don't think, at that point, I don't think Bob Noyce was head of the organization. He was just a researcher. Remember that in 1959 or before that, Ed Baldwin was general manager. Whether Bob Noyce reported to him or not, he may well have, as just an individual engineer. Baldwin was general manager of Fairchild, of the group that reported in to Fairchild Camera and Instruments. That's one question that might be explored a little further with someone like Gordon Moore, what that early organization was in '57, '58.

Lowood: It's kind of interesting to me that some of the major spin-offs from Fairchild were by people who were running the company. It raises the question of why somebody needs to spin-off into another company.

Deal: Because he noted that the eight founders made some money. Now this was important. His main reason for leaving was to see if he could make some money by starting his own company, and getting options with Rheem.

Lowood: So I suppose a plausible way to think about the early development of the R & D would have been, well, Noyce took over as general manager and his work continued, probably under Moore.

Deal: Under Gordon Moore, right.

Lowood: And then that was the nucleus of the developing R & D organization. Did Noyce play any role at all in research and development?

Deal: At the time I came, we had communication with him, he would send letters of commendation if some successful development were announced or a paper were published or so on, but we really didn't have that much direction from him, as I recall. Any direction that we got was through Gordon Moore.

Lowood: Moore was the more visible of the two, here?

Deal: Moore was R & D. There's no question that he was in cHRGe.

Lowood: You know, one talks about, say, the HP style, management by walking around. Was the Moore's style, to kind of walk around and take . . .

Deal: Gordon Moore kept track of what was going on in the facility, yes, he would be visible throughout the facility.

Lowood: Did he also hand out ideas and make suggestions about projects?

Deal: He certainly did, especially during our weekly meetings that I was involved in on the MOS development

Lowood: So he was intellectually involved with what was going on, as well?

Deal: Oh yes.

Lowood: Were there examples in the period after you arrived when he made important suggestions that pushed things in a certain direction?

Deal: No, I can't really think of any key things like that. I know that he was very active in helping make the decisions, but whether he made any key decisions himself, I'm not really sure. It seemed to me, most of those came from the people.

Lowood: Of the ones, I guess, who went to Intel, Grove was still quite active in R & D when he left?

Deal: Oh, yes. He had been made assistant director before he left, to Gordon.

Lowood: OK, I got on the subject when talking about the mask reduction technologies and all that sort of thing. So that figuring out how that fit into R & D, that was something of a legacy, I suppose, of the early integrated circuit development, and then eventually, as that progressed through R & D into a separate facility.

Deal: Right.

Lowood: One last thing about R & D, again having to do with evolutions of things through research and development to other parts of the company. Were there, were R & D labs, formal R & D labs set up in any of the production facilities outside of the Valley, the ones that were set up later?

Deal: In later years, as we went to production facilities outside California, or even in California out of the Bay Area, there were R & D groups set up. There are even to this day in those production facilities in both Puyallup, Washington, and Portland, Maine. [Puyallup was closed soon after this interview.] Portland, Maine was the first facility to produce semiconductors out of the Bay Area by a Bay Area company, and they've always had some sort of an R & D, it's more a development group. We've always tried to maintain communication with those groups, and during the years, certainly in later years, there have been what we call technology manager meetings that were sponsored by the director of R & D. So those meetings would be either held here, or at the various locations where the production facilities were established. When they were here, we would have people from here attend and review things with them. We have had the last seven or eight years, what was called the R & D or the Fairchild Technical Seminar, which I was in cHRGe of during those times. That would be a part of this meeting when we review plans for the next one each time. But we'd talk about all sorts of problems at those meetings.

Lowood: These outside of Silicon Valley labs and stuff, these would have begun, when, in the mid-seventies? Was that about the time? Or was it even later than that?

Deal: Portland started in 1962, when Charlie Sporck was manager there at that point.

Lowood: And that started as a production facility, and then whatever R & D component there was came in much later, or was it pretty much there from the beginning?

Deal: They had something there probably all the time. They had to have something that was sort of like a development group, to help establish the processes and I guess over the years it changed more into what would be called an R & D organization, rather than just a straight sustaining engineering organization. They still have the sustaining engineering, but they would have a separate R & D group.

Lowood: But these communication problems, or communication considerations that were handled through the technology manager meetings and seminars and so forth, this is something that is already beginning in the early sixties, at least?

Deal: In some way. In a less formal way, yes.

Lowood: Are there any important developments that occurred either in Puyallup or in Portland?

Deal: Well, more recently, the Puyallup group has been the strongest technically, and the 64K SRAM using bi-CMOS was developed there, in competition with another bi-CMOS process here. The one here was more technically advanced, but at the time this decision had to be made as to which one we would use, National bought Fairchild and the people who had developed the one here left. They are now still working on that in Puyallup. In fact, Puyallup has now been named a bipolar technology center for National, whereas in Santa Clara, CMOS work is going on there. So that sort of reflects the strength of the R & D group that was the bipolar group at Puyallup. [Puyallup was closed soon after this interview.]

Lowood: That actually anticipated my next question. I was going to ask if, like, say, Hewlett-Packard, their Boeblingen, their German facility specializes in the biomedical instrumentation. If Fairchild, now National, had also gone to specializing particular sites and particular technologies.

Deal: This, I think, has been the case with most of the companies that had plants outside the area. Each plant would have some sort of a specialty. For many years, South Portland, Maine, was a digital product headquarters, that's where digital products, bipolar products were produced.

Lowood: From your point of view, that's worked out pretty well, that's a pretty logical arrangement?

Deal: Yes. Seems to work out fairly well. I think the key thing, and the real problems, are to maintain the communication among those groups and back to an R & D group which is working on the more fundamental, longer-term problems. Communication is the big thing.

Lowood: Is that the way it was set up by Fairchild, that the central R & D lab, would be differentiated from the others and that it would be taking a longer term perspective, looking to future developments.

Deal: Yes, that's absolutely correct.

Lowood: And was there also, ever a formal oversight function of this lab over the others?

Deal: All the R & D managers at the plants report by dotted line to the R & D director in Palo Alto. That was the case for several years.

[END OF SIDE A, TAPE 2/BEGINNING OF SIDE B, TAPE 2]

Lowood: We've talked about the MOS projects, we've talked about the organization. One thing I wonder about, in terms of getting back to your coming into the company, into Fairchild, is whether there were other research or development areas outside of MOS that you were intimately involved in during the sixties, that perhaps we haven't mentioned yet?

Deal: Well, as I took over various responsibilities in dielectric films, first, and then materials and processes, I obviously had to get involved, at least have a better understanding and being able to direct some of these other programs in all phases of materials and processes. However, I was able, fortunately for myself, I think, to continue to direct individual projects involving surface physics and chemistry of semiconductors. I've been able to do that to this day. So on a broad picture, I was responsible for all of the development of processes and materials, but still could maintain the direction of specific projects, and I've had a series of people, generally, nearly always who would be a Ph.D., except in the first case of Maija Sklar, who was not, who would be the project leader, but I would oversee and direct the project. I would be the project manager and they would be the project leader. So I've had a series of people, about five or six over the years, that have carried out an individual project with me.

Lowood: So you've tried, in addition to moving up and sort of broadening your managerial span, you wanted to keep your hands in your core technical field, if you will, and continued to work on these individual projects?

Deal: I've been able to continue to work on them, publish papers, present papers. It's been an ideal situation in a lot of respects, and people have said they were very envious of me because I was able to do that. Typically, as you get in to management, you're not able to maintain direct project work.

Lowood: What did you do to make sure that you were able to keep that going.

Deal: I don't say I insisted on it, but I certainly suggested that I continue to do that, and it's worked out. If you're willing to do it, I think some people just aren't interested in doing it after a certain time, but I have been. I've also been able to maintain my work with places like Stanford.

Lowood: And there haven't been times, when, for example, maybe during this year you were managing the engineering group, there may have been less time, perhaps, for research?

Deal: Well, there was certainly less time during that year, although I still had an office here, and was unofficially in cHRGe of the group. They did not replace me here, during that year, because it was a temporary assignment to Mountain View.

Lowood: Just looking at the papers, it does seem to me that you've managed to continue to work pretty much on fairly similar subjects through the years.

Deal: The amazing thing is that there is as much interest today, in 1988, as there was in 1963 in oxides, and problems of oxide passivation. It's hard to believe, but we just had a symposium on that subject, The Electrochemical Society, in Atlanta, in which we had something like seventy-five papers presented, over a five-day period, on subjects that were very interesting and detailed. At the same time, they certainly indicated there are still problems and they will continue to work on these. We've got another meeting like that planned in four years. And in the 1960s, when we had all the people working in this particular area, we didn't think it would ever last more than a few years. In fact, I've been told at least twice in my career that I had better be looking for other areas, technical areas to work on, because all of the problems had been solved in the particular area that I was working on.

Lowood: You have stuck pretty much to silicon. Have you ever felt the pressure to move to some of the other, I mean, you mentioned some of the gallium arsenide things a little earlier?

Deal: No, especially since we got out of gallium arsenide at Fairchild. There was a time when we thought we might want to expand into gallium arsenide, and help the people working on those devices. And I've kept sort of informed as to what's going on there, but I have no justification, really, for working on it.

Lowood: And there hasn't been much research, then, done here on things like indium phosphide and all these other materials?

Deal: Not since we got out of that business. We included other things besides gallium arsenide in when we had the optical electronics group working on other materials. Various LEDs, light-emitting diodes, you use different materials to get different colors of lights, and there was work going on here in those other areas. But shortly after that was when they finally decided to close it out.

Lowood: Just for the record, can you mention some of the products, that were the continuations of the work that was done in the R & D groups that you were associated with, say, through the mid-seventies or so?

Deal: Well, of course, on the MOS, all of the microprocessors. There's a family that Fairchild designated FACT, Fairchild Advanced CMOS Technology, which is current today, came out of the work in our department here. Similarly, there's another product family at Fairchild known as FAST, which was Fairchild Advanced Schottky Transistors. Then another family, called Aspect, which is based on our ECL (Emitter Coupled Logic), advanced ECL devices came out of my department here. Just about everything that's in the industry has been worked on at Fairchild, and started out in the R & D lab. We were reviewing this for National, when they took over, and what some of the things were. As I said, the bi-CMOS process is another one. Even up to the past couple years, we've kept up pretty well with the developments. CCD (CHRGe Coupled Devices) originated, it turns out, originally in my department and then was separated out as a separate organization and finally is almost a separate company, now.

Lowood: So Fairchild, in contrast to some of the other companies, has stayed pretty much with developing home-grown research, in a sense? Have there been cases where Fairchild decided otherwise, that is, to go with the technology that had been developed elsewhere for a major project?

Deal: Interesting question. I'm trying to think. I can't think of any technology that we have used that was developed somewhere else. I can thing of one area that I wasn't involved in. I guess it was a sort of side type of thing, that was LCDs (Liquid Crystal Displays), for the watches. That technology was brought here from RCA, I believe.

Lowood: And that was in the consumer products?

Deal: That was in the consumer products area, right. So some of that technology was developed outside. But as far as the silicon device technology, everything that you can think of, silicon gate technology was developed here, all the interconnect technology, the device and design technology for these various products, all here. I guess where we did count on things coming in from the outside, we would hire people from some of the other areas, we would get some of it. But that was more an individual person, from whom you'd get more an individual process, or something like that. You wouldn't get the whole technology. You couldn't.

Lowood: Did you mean to say that that was sort of a planned thing, or that as people became available?

Deal: No, that was not planned, that just happened. That just happened.

Lowood: OK. For today, I'll just finish up with a couple questions about the transition in the company that began to occur, well, you've corrected me on this a little bit. You talked about GME in 1964, but I suppose it really began to accelerate with the departure of Sanders in '67, and then Noyce and Grove and Moore to found Intel in '68.

Deal: And Sporck in 1967. Because that took, the National thing really took a large amount of our production management. Those three years were the hardest on Fairchild, and probably were really the start of the problems in later years that Fairchild has had.

Lowood: It seems to me that they must have been hard for at least two or three reasons. One being just the quality of the people that left.

Deal: Intel was the worst, because they were R & D people, whereas most of the other spin-offs, they were production people, or they would be individual R & D people. Now Intel would take people for several years afterwards, one at a time.

Lowood: So Intel, after the creation of Intel, there was a continuing stream of people leaving.

Deal: A continuing stream. And yet you have to smile, then, when a few years after that, Intel would be one of the loudest complainers about people leaving Intel to form new companies.

Lowood: So key people left, and a lot from R & D. A lot of senior management slots were open. Did morale in the company fall quite a bit during those years? Did productivity drop?

Deal: Up and down. Up and down. Then you'd bring in new people, when Hogan came in he brought all the key managers from Motorola, so we were in very good shape, the morale was very high then. It probably wasn't very high at Motorola. Things looked very good. They looked good again when Schlumberger took over, interestingly enough, because Schlumberger indicated very soon that they would help support the company, especially R & D, and they did. They spent millions of dollars on bringing R & D, both from a employee point of view. This building was in very bad shape and was completely redone. New equipment was purchased, so if you can fault Schlumberger on some things, you cannot fault them on that. That probably comes in in a later discussion, but it does relate to the morale issue.

Lowood: Certainly. I was wondering, for the earlier period, with Hogan's arrival, it's not just that Hogan is arriving, but he's also taking a different title, in the sense that he wasn't just in cHRGe of semiconductor and instruments, but really of the whole company.

Deal: That was the major change in the company. They changed Fairchild Camera and Instrument really to Fairchild Semiconductor. They kept the old name, but everything else was changed. The headquarters were moved here and the whole company was run from the semiconductor end.

Lowood: So that, then, would have also been interpreted as a positive sign by the employees?

Deal: Oh, that was very positive, right. It really helped to take care of some of the previous disappointments, as far as the other people leaving.

Lowood: Was Hogan much involved with research and development? I mean, he had the background.

Deal: Very much. Oh, yes. Since he also had a Ph.D. and had been at Bell Labs and had taught at Harvard, he was very supportive of R & D, and was a very good friend of Jim Early, who was director here.

Lowood: Was Hogan himself visible here?

Deal: Yes, he was fairly visible, as I recall. He had a lot of problems, of course, to face in getting things straightened out in Fairchild, but . . .

Lowood: Do you think he was generally successful in what he set out to do with the company?

Deal: He seemed to be, although because of the downturn in business, then, in the 1970s, early 1970s, it's a little hard to know whether he was a victim of that, or whether he was responsible for part of it. I don't think anyone will really ever know that, and Corrigan was able to take over.

Lowood: Let's see. Hogan, as you mentioned, came from Motorola. But there wasn't a significant, or was there a significant influx of technology from Motorola? You mentioned sometimes when people would come from other companies . . .

Deal: A lot of the people. All the top managers came from Motorola. But not technology. R & D itself was only affected by the fact that Jim Early was brought in here from Bell. But to my knowledge, there were no Motorola people brought in to R & D.

Lowood: Maybe because of Early then, were there directions that changed as a result of things that were brought from Bell at that time?

Deal: I don't recall. The only thing that happened, as I said, is that after a year or so, why, we were drastically reduced, and that was the negative direction.

Lowood: That was under Corrigan?

Deal: But Jim Early did not bring anyone else in. He kept the people that were here.

Lowood: Can you state an impression, compare Hogan in any way, in terms of his impact on the company, to, say, Noyce? Was there a big change in the way the company was run at that point?

Deal: Not as far as R & D was concerned, as I recall. I don't recall much of a change.

Lowood: As you say, the R & D philosophy didn't change.

Deal: No. The organizations didn't change that much, any more than they normally would have, I think, anyway. And the projects we worked on were pretty much the same.

Lowood: The company expanded up until '71?

Deal: '71 was the drop.

Lowood: Up until that point, though, there had been a fairly steady expansion in the size of the company?

Deal: Well, in R & D, in 1966, the R & D building, as represented by the first two wings of this building, was completely full, and so we then, in 1966-67 built the third wing, which had a second story, and doubled the size of R & D. That was quite an expansion.

Lowood: I noticed from the chart, I don't know if I mentioned this already, that there were 541 employees, 210,000 square feet. That's really quite impressive for an R & D operation..

Deal: That's large, that's the size of this building, 200,000 square feet.

Lowood: What happened in 1971 in the industry that caused the downturn?

Deal: It was one of the many so-called recessions. I can't recall if there was a general recession in the country at that time or not. There may have been. There certainly was in the semiconductor industry.

Lowood: Prices fell for the bread-and-butter items?

Deal: There was no market, I think. There wasn't any business. Now I don't think it was probably any worse than a couple of other times, but it was as far as R & D was concerned. I think in other times we had similar recessions, I can't really remember the dates at this point, but R & D was not affected, and it was in this one particular instance.

Lowood: Were the markets, up until that point, and I'm talking about the sixties?

Deal: There was a recession, as I recall, in the middle-late sixties, as well.

Lowood: I see. So it was beginning to become cyclic at that point.

Deal: Right.

Lowood: Was the market for this company primarily industry, government, military?

Deal: We've always been a fairly strong military company. We had quite a bit of business with the military, but we've been very strong in computers, because of the bipolar, high speed bipolar parts. Fairly broad based, because we had such a broad-based product line, all types of products were represented: linear, analog, discretes, logic, memory, pretty well spread across the entire spectrum.

Lowood: The military market would have involved hardened things?

Deal: Right. We had various contracts. In fact, we had a few contracts here in R & D on both hardened device, primarily bipolar in those days, and also MNOS, non-volatile memory, some other things. That all stopped with Schlumberger, but pre-Schlumberger .

Lowood: Because it was foreign-owned?

Deal: That was part of it. Schlumberger also didn't want it. We were involved in some contracts, here, in R & D, that were strictly research, but they stopped those, because they didn't want to be involved with the government whatsoever, and it was not a problem as far as being foreign-owned. It was their desire. In fact, some of the product lines we continued.

Lowood: Was there DOD funding for research that was not connected with products? I'm thinking like maybe from ARPA or something like that.

Deal: Yes, I had some myself that were strictly R & D.

Lowood: What kind of areas were those in?

Deal: The ones I had were related to MOS and oxides, and I was doing cooperative work with government scientists, on which we then published joint papers, which we included in the list here. We were involved in the DARPA support of the Stanford SUPREM program for many years. I have been working on that program, both with and without support, since 1975.

Lowood: I'll ask more about that when we get up into it, into the seventies.

Deal: On the other hand, the philosophy of Fairchild was pretty much that we would not be heavily supported by R & D programs, and I think at any given time we didn't have more that five to ten percent of support. And the only exception to that was the CCD effort. The CCD effort was typically supported 100% by military contracts, and still is. There was a Fairchild space and defense group, part of Fairchild Camera, and that was the one group that did depend primarily on the government contracts. And then they became a part of that later, and they are a part of it today, as a center of Schlumberger company.

Lowood: Why was the military particularly interested in the cHRGe coupled devices?

Deal: Because it's an imaging device, can be used in surveillance and things like this, and in aerial photography, which of course Fairchild was always interested from the start. Fairchild Camera and Instrument got its start in it. But CCD devices were also, you can view at night using infrared. There are a number of military applications for solid state imaging.

Lowood: But the funding that came here was strictly for research and development in the general area, not for specific?

Deal: Well, they also sold devices, and then, together with the space and defense group provided solid state cameras, as I understand it. I'm not too sure what they're doing today.

Lowood: Was there ever any kind of, now I'm speaking just about the R & D lab here, was there ever, were there projects that required strict oversight different from the regular monitoring of people inside of the lab.

Deal: Yes. We've had pseudo-secret clearance projects, but not a lot. Some of the radiation work we've had here, we've had more control over. I think, for instance, people working on the projects had to have a clearance classification, and probably had to be citizens.

Lowood: And some of these already occurred during the sixties?

Deal: Yes, I think they started in the 1960s, late sixties.

Lowood: And I guess Fairchild as a whole would have been involved in aerospace, at least going beyond the semiconductor part of Fairchild, the instrument division probably was involved in a lot of parts of aerospace work.

Deal: Right.

Lowood: OK. I think this might be a good place to stop, then.

Deal: I guess we've talked enough about colleagues.

Lowood: Yes, unless, we have a few minutes here, if there was anyone. I've written down on the outline, we have here C.T. Sah, Snow, Noyce, Moore, Grove, Sanders.

Deal: The only thing would be, whether you're interested in knowing where these people went and what they're doing.

Lowood: Well, if you wouldn't mind going down the list and just telling that, that might be interesting.

Deal: As I mentioned, C.T. Sah went to the University of Illinois, shortly after, about a year or so after I got here, and he's been there ever since, as a professor of electrical engineering. He's the only person at Illinois working on silicon devices. Everyone else is on gallium arsenide compound semiconductors.

Ed Snow left in 1968 or '69 to help form a company called Reticon, which was in imaging. They were later bought out by some other company, I think it's EGG, something like that.

Of course, Bob Noyce and Gordon Moore and Andy Grove all went to Intel. Andy Grove changed completely from a researcher at Fairchild to a tough production boss at Intel, and also writing of books and gossip columns.
I didn't have much contact at all with Jerry Sanders. I don't recall that at Fairchild he had that high a position in the marketing organization. He may have ended up as marketing manager, general manager.

Charlie Sporck, of course, was production manager and general manager here, and he's my boss today. And I think he's a lot milder and more supportive in R & D today than he ever was in the early years at National and at Fairchild. He's supported such things as Sematech and SRC. He supported our transfer of several of our programs with universities that Fairchild had and National did not have. Very supportive, and saw that we got to keep those programs going, something that National had not done at all.

Lowood: As I look at these six or seven names now, generally speaking, probably with the exception of Sanders, these are people generally that don't fit into this image of the flamboyant kind of manager.

Deal: No, they're the scientific, no, certainly not. Well, Andy Grove, you might . . . He's a hard one.

Lowood: He's moved in that direction a little bit. I see his column in the Sunday paper now on how to handle management problems. It's hard to imagine it's the same person that was working on some of these things twenty-five years ago. Were there any other people that you mentioned in the course of these two hours that you think would be worth mentioning, in terms of where they ended up? I'll just quickly glance here. I guess that covers pretty much the people that you've collaborated with.

Deal: Charlie Bittmann, who was instrumental in getting me here, went to HP, and he's now retired.

Lowood: And you mentioned that Maija Sklar had . . .

Deal: And Maija Sklar has retired from Fairchild.

Lowood: Jim Early, likewise.

Deal: Jim Early is now retired, and a number of the earlier Fairchild people that stayed on at Fairchild have pretty much retired.

Lowood: There was one I was curious about, who headed basic physics for a while. His name was Leifer.

Deal: I have no idea where he went. I think the person that was here after him, though, was Herb Kroemer, and Herb Kroemer went to the University of Colorado at Boulder and now is at Santa Barbara, and he's quite well known in the field of compound semiconductors, and I believe he took over that group.

Lowood: OK, so we'll pick up then with the facilities and move through the seventies and eighties, and then I'm positive that we'll finish up. It probably won't even be two hours.

[END OF INTERVIEW]