Oral-History:James Gibbons

From ETHW

About James Gibbons

1726 - gibbons.jpg

Gibbons studied engineering at Northwestern, graduating in 1953. He got his PhD at Stanford, and became a professor there soon thereafter. After a brief period working with William Shockley, at Stanford’s direction, he set up a solid-state electronics lab at Stanford. The work at this lab spawned the graphics chip for SGI and the MIPS chip for risk processors. After setting up the lab, Gibbons moved over to work on the physics of fabrication—wafer fab technology, with epitaxy and ion implantation.

About the Interview

JAMES GIBBONS: An Interview Conducted by David Morton, IEEE History Center, 31 May 2000

Interview #399 for the IEEE History Center, The Institute of Electrical and Electronics Engineering, Inc.

Copyright Statement

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

Request for permission to quote for publication should be addressed to the IEEE History Center Oral History Program, IEEE History Center, 445 Hoes Lane, Piscataway, NJ 08854 USA or ieee-history@ieee.org. It should include identification of the specific passages to be quoted, anticipated use of the passages, and identification of the user.

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

James Gibbons, an oral history conducted in 2000 by David Morton, IEEE History Center, Piscataway, NJ, USA.

Interview

Interview: James Gibbons

Interviewer: David Morton

Date: 31 May 2000

Place: Palo Alto, CA

Family, childhood, and educational background

Morton:

Tell me about your background and about things that are not on your curriculum vitae. Where and when you were you born? Also, how did you become an engineer? What early experiences led you to become an engineer?

Gibbons:

I was born in Leavenworth, Kansas. My father was a guard at a prison called Fort Leavenworth. The Fort had been built by a man named Henry Leavenworth who had been a Colonel in the United States Army. It was the most secure stockade west of the Mississippi from the Indian-fighting days.

Morton:

I just visited Alcatraz. “The Birdman of Alcatraz” was really the “Bird Doctor of Leavenworth.”

Gibbons:


Audio File
MP3 Audio
(399 - gibbons - clip 1.mp3)

Correct. Since Fort Leavenworth was a secure stockade, the Federal Government rented it; they loaned it to the Department of Justice. It became one of the maximum security penitentiaries in the country, second only to Alcatraz. When Alcatraz was in operation, you could not be paroled from Alcatraz. You could be paroled from Leavenworth. In fact, my father’s first job when he signed on at Leavenworth was manning a Thompson submachine gun across from Al Capone’s cell. Capone was on his way from Chicago to Alcatraz.

I was born at Fort Leavenworth and lived there until I was eight years old. Then the Department of the Army needed the Fort back in order to secure POWs. In the interim, they had built a prison up the Missouri River, which is the Leavenworth prison that everyone knows about. All the guards were moved.

Because of the fraternalistic connection between Fort Leavenworth and Alcatraz, it was a matter of honor amongst the guards at Leavenworth that they would apply to work at Alcatraz. My father applied. My mother hoped that he would not be hired because the families of the guards lived on the compound, typically. I could have moved at age eight to Alcatraz. He didn’t get the job. They sent him to a minimum security penitentiary in Texarkana, Texas, which is where I lived from age eight to age sixteen when I went away to college.

The difference between a maximum and a minimum security penitentiary in the federal system is huge. First of all, most of the violent crimes are not federal crimes, they are state crimes. So Sing-Sing or San Quentin are likely to have people who are more instantly violent. Al Capone was caught for tax evasion. Minimum security penitentiaries have people like E. Howard Hunt and Gordon Liddy in them. Texarkana was a minimum security penitentiary that held people who were convicted of mail fraud—opening mail and removing money from it. Also, the inmates at the minimum security penitentiaries do things: they work a farm, or they made brushes, or they built the houses that are out on the prison compound.

We were among the first families to move out to Texarkana. As an eight year old, there was nothing to do except to watch the inmates construct houses and gradually build friendships with the inmates. Later on I played baseball with them and so on. So I developed some friendships with these guys. Today, I have a company that teaches juvenile offenders how to manage their anger and to walk away from fights. We have been so extremely successful with the program that we are now in six counties in California in the alternative school system in Los Angeles County in all sixty schools. We are also in places across the country, in both regular schools and court-appointed.

You learn different things when you grow up with people. To some degree (and it’s an exaggeration for sure), in a way I was kind of their kid too. When I would slide into second base, an inmate would make sure that I knew that he had tagged me, even though I might be safe. He would say, “Just in case you are ever lucky enough to make it to the major leagues, I want you to remember this day.” These sorts of experiences happened to the half a dozen of us children who happened to grow up there and played baseball with the inmates on the reservation.

East Texas is an unusual place to grow up during the period that I was there. A high school classmate that was ahead of me who everyone would instantly know is Ross Perot. I was in Ross Perot’s class for trigonometry because if you were in what would today be called an AP class, you were accelerated to the next year. I had also been skipped a year earlier. The mathematics and English classes were accelerated, so I was in some of his classes.

In this high school, there was very good instruction in English: two Shakespeare plays per year, a lot of writing experience, and a lot of poetry. When the good students left high school and went to college, they aced out of freshman English.

The kind of advice you would get about what to do for a living was entirely by indirection. I happened to be a musician. I thought I wanted to be a jazz trombonist. I was good in math, I was good in science, I was good in English, I played trombone, and I played baseball with the inmates and with the team.

Texarkana was a class-D Yankee farm club, which meant Pro-Am, basically, so you could play occasionally with them as an amateur when the guys were trying to make the big leagues. We never thought we were trying to make the big leagues. The climax of my baseball experience was that I was a good hitter and a good fielder. They figured they would take a few kids who might be major league prospects to watch somebody that they weren’t quite sure was going to make it, but they thought he might. So they loaded us all on a bus and took us a hundred miles or so to a little town called Commerce, Oklahoma, to watch this kid do his batting practice. There was a right-handed pitcher, and every pitch he just creamed, most of them sailing outside of anything. Then they brought in a left-hand pitcher, and this kid goes to the right-hand side of the plate and he does the same thing. It was Mickey Mantle. They didn’t know he was going to make it. Those of us who thought that we might be able to play this game thought, “Hey, this is a good game. I can play this game. Yeah, but not the way he does!” So I quickly decided that music was a better thing for me to do.My high school advisors were telling me that I should be an engineer because I was good in math and in physics. The practical thing to do was to be an engineer. I did not know the difference, and no one else knew the difference. Today, I would be a mathematician or a physicist. Back then, it was go be an engineer. But I wanted to go to a school where I could play music, because I thought, “Do I really want to be an engineer? Or do I want to be a musician?”

Undergraduate engineering studies and laboratory employment

Gibbons:

I applied to various universities. I had four-year scholarships to schools in Texas. I also had a music scholarship to Louisiana State University, which was a five-year scholarship program through a master’s degree. But southern schools at that time were a year behind really good engineering colleges. If you went to the University of Texas, you took college trig and algebra the first year. If you went to Northwestern (where I happened to go), you started with geometry and calculus, which was the second year in Texas schools. So the south was behind by at least a year. I was not sure that I wanted to become an engineer anyway. I had a partial scholarship to Northwestern, and thought, “This is great. I can play in Chicago and see whether I like this.” So I went to Northwestern and signed up as an engineering major. I did what my Texarkana high school was happy for me to do, which is I aced out of freshman English.

I played jazz for a couple of years. I combined trying to be a musician with trying to be an engineer, until I figured out that I loved to play jazz. However, you find that when you are playing with real professionals that it is not a life that you can necessarily see yourself leading for a long time. If you have any idea about having a family and any kind of family integrity, you see that it’s very difficult to do in these circumstances. You’re on the road all the time. You’re playing when everyone else is working and you’re working when everybody else is playing. It’s hard. A lot of divorces. A lot of drugs. Here I was, coming from this protected east Texas cocoon into an environment that is something that I had never imagined. I went from Texarkana, Texas, to the streets of Chicago.

I eventually decided that I didn’t want to do it anymore, so I stopped playing. I play now, but I didn’t play for twenty years. When I picked the trombone up again, I played in a small brass quintet that plays Bach and Handel and so on, not jazz.

So I became an engineer by indirection. First I thought I would be a baseball player, and then a musician. I happened to be good in the subjects that east Texas high school people thought, “You go be an electrical engineer since you’re good in math and physics.”

Electrical engineering always had—and I think it’s still true—the most demanding curriculum. It required people who were very good at both physics and mathematics, and an ability to think analytically as well as be creative about physics. So it was good advice I received in high school, even though it was the average kind of advice. I am indebted to my high school. I think I received a good education there because I could do different things and they aimed me in the right direction.

When I enrolled at Northwestern, they had a program that turned out to be very important for me. Northwestern is a North Shore school, mostly kids who came from well-to-do families. My mother and father put me on the train. My mother had sewn four hundred dollars into the bottom of my pocket. That was my ante for college. It paid sixty percent of the first year’s tuition. I had that and my trombone, and a half-time scholarship. I made it through.

One of the things that was helpful financially—but more than that, intellectually in terms of figuring out what engineering was like—was that at Northwestern you had to do co-op. It was not an option then. It is now. That is one of the reasons that I selected going there. Not only could I play in Chicago, but I also realized that this was a way for me to work part-time, so I would be able to generate the income that I would need to pay for the next quarter. After the first year, the class was split in half and half went to co-op and the other half stayed in school. Then they reversed so that the employer always had somebody.

Morton:

Where did you co-op?

Gibbons:

I did my co-op at a company called Tungstal. Its headquarters are in Bloomfield, New Jersey. They made vacuum tubes. But they had a lab in Chicago on Grand Avenue, right in the area where I used to play trombone. The beauty of the Chicago lab was that all the television manufacturing was done in and around Chicago at that time: Admiral, Motorola, Zenith, et cetera. All the major tube companies had their facilities on the East Coast, but they had labs in Chicago.

Tungstal had a lab in Chicago that was managed by a man named Ed Atkins, who had never attended college. He had twenty-nine patents when I started working for him at the co-op job. I didn’t know which end of a soldering iron was hot. I knew nothing about the profession. I knew how to play trombone and I could recite poetry. He loved it. He thought, “Here’s this untouched guy, and I am going to help him.” He was skeptical about a university education, as inventors often are. He had basic patents on automatic volume control and basic patents on carburetion. Just a brilliant, brilliant man.

I worked for the company five days a week and I also came in to work for Ed Atkins on Saturday to work on his inventions. He would come in with the Saturday Review of Literature underneath his arm, which he had partly read coming to work on Saturday and going back. A very sophisticated, urbane, very thoughtful man. If he would ask, “Why is something so?” and I would give a mathematical explanation of it, he would say, “That’s bullshit, Jim. You didn’t tell me why. You just recited some crap. Why is this the case?” So I got accustomed to thinking about things in very fundamental terms that you could explain essentially to an intelligent person. An inventive, creative, intelligent person, but nonetheless untutored in the profession.

It was like learning how to play ball with the inmates or learning how to build houses with inmates. A lot of stuff that people know they did not learn in the normal way, and yet they are able to do things that are creative. They are artistic. There is an art to this profession that gets lost sometimes if you are too mathematical. You can be creative with modern tools, but the creativity becomes focused in a certain way because of the nature of the tool. With fewer tools around, you had to imagine many things. It was an interesting way for me to grow up.

I had a job that was perfect. I was working with an ideal person. I had a chance to work on all of his patents. If we went to Motorola it was because they had a problem that they couldn’t solve. I had a chance to work on really first-class problems. It was interesting. It was a consulting lab essentially, maintained by the company on behalf of the television tube industry. We had sold twelve percent of a full complement of tubes to Motorola. We were twelve percent of Motorola’s television business. It was a big deal. So great job.

National Science Foundation Fellowship and Hewlett Packard employment

Morton:

Was this sometime during the early 1950s?

Gibbons:


Audio File
MP3 Audio
(399 - gibbons - clip 2.mp3)

Early 1950s. I finished Northwestern in 1953. It was a five-year program because of the co-op. You are spending half of your time after the first year not in school. I finished the program, and I won a National Science Foundation Fellowship.

My family lived in Tennessee between Nashville and Clarksville. I am related to everybody in any graveyard there. My family was all hill country farmers, scratching a living from mostly granite land covered with a little bit of topsoil. My father was a schoolteacher. He taught in a one-room schoolhouse, except when the needs of the farm were more important, then all the children who were supposed to be in school were outside doing what they had to do. In 1931, the state of Tennessee decided to reduce the salaries of their teachers. They reduced my father’s salary from fifty dollars a month to thirty-five dollars a month. I was not yet born; I have an older brother. My father said, “I cannot support my family on this.” So he applied for a job in the civil service, thinking, “I will get into the Post Office in Nashville.” In those days, if you did well on the test, which he did, they took the top person and clipped his score and resume to the next job. The next job was in Leavenworth, and they asked, “Do you want it or not?” He said, “Sure.” They asked, “Can you shoot a gun?” They’re talking to a Tennessee farm boy. He said, “Yes, I can shoot a gun.” Otherwise, I probably would have gone to Vanderbilt. Or if we had gone to Alcatraz, maybe I would have gone to Stanford.

So, that gets me through my undergraduate years. You must understand why I interpolated that story. I had received a job offer from RCA upon graduating. My father’s advice was to just take the job and start working. Thirty-five years later you could retire. That is what you did. That’s why I went to college—to get a good job.

I had a National Science Foundation Scholarship for one year. That’s what you received in those days. I could go anywhere with it. I had an advisor who said, “You really need to have at least a master’s degree. You should do a Ph.D., but you should have at least a master’s degree.” I applied to MIT, Caltech, and Stanford. I had offers elsewhere, but those were the schools in which I was interested. So I asked him, “What should I do?” He said, “Caltech isn’t really an engineering school. It is a science school.” This is him talking. Whatever the reality is or was is something else maybe. He said, “You can go there and you will do fine, but I don’t think that’s the right place for you. MIT is too traditional in the way they do things.” Then he came to Stanford, and a big smile came over his face. Remember, this is 1953. He said, “Stanford probably isn’t as good in terms of the quality of the faculty as MIT and Caltech, but there is something about the place that is different, and I think you will like it.” I asked him what it was. Today he would have said, “Well, it is entrepreneurial.”

In 1953, Hewlett Packard had been in existence for about seventeen years. Terman was interested in building something. There was a certain degree of amused forbearance in the rest of the world. HP was fairly small. What were they trying to do when there was General Radio and so forth? MIT was the epitome of everything. But there was something about Stanford that my advisor thought I would enjoy. So I went there. I finished my master’s degree and received an extension of the NSF grant for a couple of years. You had to take the test every year then. Nowadays you don’t; you get three years at a pop. Back then every year you would re-compete. It was a little bit uncertain.

In the summer, I worked for HP because Barney Oliver had given some lectures at Stanford. Something just happens and you think, “Hey, that’s a neat deal.” He gave two lectures and a course for all the master degree students. He was giving a broad description of Foray transforms and how you use them in various communication problems. I was thrilled by this. I thought, “If there is any way that I could work for him, it would be fabulous.”

I needed a summer job, so I applied at HP. They asked, “What is it that caused you to come here and apply?” I said, “Barney Oliver gave these lectures and I am really interested. If there is anything that I could do in that area, I would be happy to work for you.” They hired me for the summer and put me on the design of a particular power supply that they wanted to build as a workhorse for clash-drawn  tubes.

Morton:

Did you come out of Northwestern with a particular focus on electronics?

Gibbons:

Yes. The transistor had just entered the curriculum the year that I was leaving, or a couple of years before that depending on where it was. At Tungstal we were looking at transistors. People thought that transistors would never replace vacuum tubes, but we were just fooling around with them. Ed Atkins said, “Let’s work with some of this.” One of the things I did at Tungstal was to build a transistor tester that would measure transistor parameters and tell you about temperature dependence and so forth that no one had done before. Ed Atkins’ view of what to do was not to develop the scientific underpinnings of it and create a great machine, you do something like that and then write it up in Radio News. So my first publication was in Radio News while I was still an undergraduate, and it was on my transistor tester. There were people at Bell Labs, of course, who saw it and said, “Ah, that is a terrible way to do this.” (I found this out later when I worked at Bell Labs.) But from Ed Atkins’ point of view, it was a very creative way—a very simple thing that anybody could do. You could appreciate exactly what was going on. It gave you first order of information about how the transistor was going to work. That was what he was interested in. I built it for him and he said, “Here, write this paper. It will make you famous.” Well, it made me famous in the Radio News circles and it made me infamous elsewhere because I didn’t make a good scientific project out of it.

It was interesting because it helped me see the degree to which sophistication also carries with it some other consequences and nuances that separates a sophisticated approach from a hands-on approach, more embarrassed kind of an approach to things, but something that still has a lot of insight in it. Inventors do not tend to be sophisticated. They are very sophisticated in the way that they think about things and in their ability to imagine things. But the academy is one thing and the inventors are something else. I have my feet in both of those. I worked for an inventor. Atkins was driving me to be an inventor. I was thrilled to be able to make inventions, and I was trying to do something respectable on the other side of the street. It was an interesting way for me to grow up.

I came to Stanford. For the first quarter of my master’s degree, I looked at the courses they had. The education that I had received at Northwestern including the co-op experience persuaded me that I didn’t need to take the electronics class at Stanford; I had already done all of that. I didn’t make friends with the electronics faculty because I went into my advisor and he wanted me to take the standard curriculum that everyone else takes, sign my card, and get me out of there. I said, “I have already done some of this. I want to do something else.” I wanted more physics because I wanted more depth than the typical EE master’s degree student who came to Stanford and then went out somewhere (not Silicon Valley) and find a job.

So I am again sort of at odds with the academic environment because I do not want to do what they want me to do. It was almost like going to Northwestern. I did say that I aced out of freshman English. The problem was if you were in the School of Engineering, they had your entire curriculum plotted out for you, and you took sophomore English in the School of Engineering with six other people who happened to get through freshman English. You had an instructor, whereas the man who was teaching the other sophomores was Bergen Evans, who was a Shakespeare authority and also a well-known literary talk show host. Everybody wanted to take his class, including me, so I had to screw up my schedule to take his class. I did not make friends with the School of Engineering for doing it, but I was happy that I did it. I thought, “Look, I am coming here to get the right education, and that does not include taking subjects that I already know.”

My first summer at Stanford I worked at HP. I started working on this stuff. I had everything working, but in little pieces—I had the amplifier working here and several other parts working over there. The feedback loops required in a well-regulated power supply are such that the components need to be close to each other, and also shielded; otherwise they will pick up stray signals, create oscillations, and nothing will ever work. It is difficult to build one piece on one part of a bench and another piece on another part of a bench and then try to connect them. I could make this work or that work. The bench had other items on it that I was not allowed to move, but nobody had told me why. They said, “You can work at this bench. You’re a summer student. That’s it. Do what you can.”

I was within three weeks of returning to Stanford for my second year, and I had been getting increasingly irritated about the fact that I could not move anything. Someone came up behind me and said, “You haven’t moved anything, have you?” And I thought, “Not another one.” So, I turned around and said, “No. I haven’t moved anything. But if I can’t move some of this stuff, I won’t be able to finish this project!” The man holds up his hands and he says, “I am Bill Hewlett. I like your style.” Everybody is standing around. This must have been plotted out in advance. People were standing on their desk. They know what is going to happen. The whole lab dissolves. There were only thirteen engineers in the R and D labs at HP at that time. I was the thirteenth, a summer student. He said, “By the way, Jim, you are probably wondering why the bench is kind of damp when you come here in the morning, aren’t you?” I said, “Yes, sir.” I had wondered. Here is this pile of crap in the middle of it with wires coming out of it. He came in before anyone got there and sprayed it because he was trying to do a study. He had built a simulated haystack for a farmer friend of his, and he was trying to figure out what the conditions for spontaneous combustion were going to be so that he could help his farmer friend make some measurements and say, “You are going to have to spray down your hay now.” That is what he was doing every morning before I got there. He said, “I now know what I need to know,” and he took all of his stuff off and I had a full bench to work with and I actually got the project bolted together. We laughed about that much later when I became the Dean of Engineering.

It was a great place to work for exactly the same reason as Tungstal had been: They wanted to build high quality products, they had an inventive kind of spirit to them. My advisor was right to send me here, and I was lucky to get this job at HP.

Ph.D. studies and Bell Labs employment

Gibbons:


Audio File
MP3 Audio
(399 - gibbons - clip 3.mp3)

When I came back to Stanford to do my Ph.D., I had had a fabulous set of experiences in the way that I grew up with the constant juxtaposition between the real world and the sophisticated academic way of thinking about the real world. I can’t imagine having done many of the things that I have done without having had a solid foot in both of those camps.

I completed my Ph.D. I went to Bell Labs. I started a research project. I had a Fulbright Fellowship. I was at Bell for three or four months so that I was could be on the payroll. I then took a leave of absence that had been arranged in advance so that I could go to Cambridge Trinity College to do my Fulbright.

While I was there, transistors were still essentially germanium and were used in little pocket radios and so forth, but still kind of low-end consumer electronics. There were university programs in transistor physics, but nobody built transistors. The view at universities at that time was that you treated transistors the way you did vacuum tubes. Universities did not build vacuum tubes. Occasionally there might be someplace that did. But as a rule, you bought vacuum tubes and you designed circuits with them. You bought transistors and you designed circuits with them. That is what I did my Ph.D. thesis on, and that is what I worked on at Tungstal all those years before. You took a transistor, you measured its parameters, and then you went off and designed things with it. So I was a circuit designer, but with a strong physics orientation.

I went to Cambridge and worked on building magnetic resonance imaging kind of things, magnetic detectors. While I was there, Bill Shockley moved his company to Palo Alto in a Quonset hut that is next to Sear’s on the San Antonio Road, with the express purpose of beginning to build commercial devices in silicon. Crystals were a half inch in diameter and six or eight inches long. Shockley wanted to be end to end, so we bought silicon; grew crystals; sliced, lathed, and polished the crystals; diffused impurities into them; and built devices. The device upon which the company’s first commercial success was hopefully going to be based was on a PNP diode. It was a switch for a telephone. If you had two lines that you wanted to connect, a subscriber to a subscriber, you had a switch on which you put on a ringing pulse, a big voltage pulse. The diode turns on and stays on while the conversation is going on. When the conversation stops, you shut it off.

Morton:

This sounds like something he brought with him from Bell Labs.

Gibbons:

Yes. It was a Bell Labs invention, and we were a Bell Labs licensee. John Lindville, who was my thesis advisor, came out here from Bell Labs to start the transistor electronics program. But again, it was you buy the transistors and you build devices with them. John, Fred Terman, and Bill Shockley over a period of time decided that perhaps they could make silicon devices at Stanford. They set up a program that would allow faculty and students to actually build devices on the theory that it would be very different from vacuum tubes. You would make devices and then you could study and vary the properties, and you could not do that with vacuum tubes. You could not get in there and pinch three grid wires closer together. You could, but somebody else had to do it for you. So this was a hands-on proposition that they had in mind. They wrote a letter asking me to come. It was an entrepreneurial undertaking. It is what Stanford would do. “We don’t know how to do this, but we want to set up something. There is a guy down here that does know.”

They hired me to come and take a faculty job. I was immediately put to work half-time with Shockley. I started working for Shockley just a few weeks before the Fairchild Eight left. John and Fred Terman had worked out an agreement with Shockley that I would work for him for two years to learn how to do everything. But while I was doing it, I would be half-time there, then I would come to Stanford and teach classes and I would build a laboratory here. So I started the Solid State Electronics effort at Stanford, a new, young assistant professor learning how to do this.

It was an experiment to see if we could actually transfer technology from industry to the university. It turned out that six months from a cold start we had really good devices—much faster than anybody expected. So the ONR contract that we had for supporting this device was suddenly doubled. The contact amount was doubled and we were asked for three years worth. We just got off to a booming start.

Morton:

What exactly were you making?

Gibbons:

We were making a little PNP diode. It was going to be used to correct the temperature characteristics of an avalanche diode, which were used as a voltage reference in a power supply. This was a very sophisticated device and it made a beautiful voltage reference. That is what we first made. We published some papers and did the right science. We built the lab and got this enterprise going in six months.

When I was trying to decide whether to do this, I received a letter from Shockley. He said, “I want to explain to you what I expect of you if you are going to take this job. I have a lot of very smart guys here, but they seem to think they are on a post-doc instead of working to make these devices and get them out there. Here is the way I want to work. I will assign the problem. I will tell you the way I think you need to work on it. You can disagree with that if you want. You can tell me how you are going to work. But I will select which way we are going to go, and you are going to go do that.” So I wrote back what I would have said to anybody, “I will put my shoulder to the wheel. If I am not a satisfactory employee, I presume you will fire me.” He wrote back saying that was an adequate answer.

So I showed up, and the first day I am ushered into Shockley’s office immediately. He said, “I want to give you a little test.” I replied, “Yes, sir.” I am a very pliant kind of guy. I remember this very clearly. He had a stopwatch in his hand, and he said, “One hundred and twenty-seven people enter a tennis elimination tournament. That means that in the first draw, since it is an odd number, only one hundred and twenty-six will play; somebody is going to draw a by. That will result in sixty-three matches. The next time there will be sixty-three people plus the by, so there will be sixty-four players and you will then have thirty-two matches in that round. So how many matches do you have to play to determine a winner?” Well, I just happened to think about the problem in a different way. So fairly quickly, before he even punched his stopwatch to start the time I said, “It must be one hundred and twenty-six.” And he said, “What?!” I said, “I believe it is a hundred and twenty-six.” He barked, “How did you do that?” I said, “Well, there is only one winner, and it takes a match to eliminate somebody. So if you have one hundred and twenty-seven people, you must have to eliminate one hundred and twenty-six. So you have to play one hundred and twenty-six matches.” He exclaimed, “That’s the way I do it!”

Shockley then starts another problem, which is actually simpler than the first one, but by then I am just wiped out. So he starts his stopwatch. After a while I can see that he is beginning to relax. Finally he stops the watch and he says, “Jim, you’re at twice the lab average on this. Do you want me to tell you the answer, or do you want to keep working?” I said, “It is your call, sir.” He explained it, and I thought, “Yes, right.” Then he said, “Okay. Now you can go. Here is the man you will be working with. Here is the project. Here is what I want you to do.” I leave his office and there is Bob Noyce waiting for me. Everybody in the lab asked, “So, how did you like your test?”

I began to work for him, I learned the process, and I transferred up here. I was designing circuits for those devices while I was trying to learn things. Several weeks afterward, Noyce came to me and said, “Eight of us are leaving and we are going to form a company. We would be happy to have you join us.” I said, “Oh no, I don’t want to do that. I came here to learn how to do this stuff, and I am going to set up a lab at Stanford. I want to be a faculty member.” He said, “Okay.” They did what they said they were going to do and I did what I said I was going to do. It was probably the most expensive decision I ever made.

Solid-state electronics lab at Stanford; lab producing graphics chip for SGI and MIPS chip for risk processors

Gibbons:

But we did start this lab. We brought people into the lab who were much better known than I was: Gerald Pearson and John Moll and so on. Once we proved that we could do this, all of a sudden Stanford had done something that nobody else had done—the integration of solid state physics. We became interested in problems that had enormous academic opportunity in them because we had a lab where we could do things that nobody else could do. When we started, everybody said, “This is crazy. Nobody is going to do that kind of stuff. Students aren’t going to be able to find jobs. GE is going to be the transistor manufacturer because they are the vacuum tube manufacturer—GE and Philco and et cetera. They are going to be doing this; they are already making transistors. What do you think you are going to do?”

That was in 1957. I am going to fast-forward to this lab, which we started in 1980. The Center for Integrated Circuits was founded years later to study problems that lay at the interface between hardware and software. When would you elect to build a chip to put everything into a chip so that everything is hardwired basically, versus when would you say most of this is going to be done with software and a mainframe? Those were the choices you had. Either hardwire everything or do it in a mainframe. Those were two extremes, and there were many problems in the middle where you thought, “That can’t be the right combination of software and hardware.” This was essentially at a point where you might think of microprocessors as really just controllers; they were not thought of in the way that they are today.

So we built this lab. The things that came out of it were the graphics chip for SGI and the MIPS chip for risk processors. Sun and Silicon Graphics and Cisco were born within a few hundred yards of where you are sitting. Some of them came out of this lab here. The first MIPS chip came out of that lab.

To get to that lab, you had to have started twenty-three years earlier by saying, “We are going to be in silicon processing technology.” Because Stanford did it, we started attracting people who wanted to work with graduate students who were going to be doing Ph.D.s in this area. Other places kept thinking, “Nah, it will never work.” Once we proved it worked, then we got into technologies. My career after building the Solid State Lab switched into the physics of fabrication—wafer fab technology. After working with epitaxy, I started working on ion implantation. Again, we thought it was very interesting and we did a lot of work in it.

PN junctions and ion implantation

Morton:

When was that?

Gibbons:


Audio File
MP3 Audio
(399 - gibbons - clip 4.mp3)

I started working on ion implantation in 1963 at Stanford. I went on sabbatical to Denmark in 1963 to 1964. I went to Denmark because a scientist from the University of Copenhagen had been here the year before. I thought that Denmark would be an interesting place. Danes like Americans and so forth. Ever since Niels Bohr, the Danish scientific establishment had been sold on the idea that science is good for Denmark. So they were spending an enormous amount of their country’s budget on science that they could never use. There was no earthly value for Denmark to have this huge nuclear physics establishment. The physics professors were Bohr’s students. One of those professors was in a place called Aarhus.

So they had built an ion accelerator out in Aarhus. A parliament representative had asked what the practical implications of the accelerator might be, which was an unusual question to come from the funding agency. Because this is a “you do science for science’s sake” deal in Denmark.

Neils Meyer showed up and said, “Jim, we have this ion accelerator. Do you think you could do anything with it?” I said, “Can I get a beam of phosphorous on it?” He said, “I don’t know. I will call.” We were standing at the airport. Meyer was calling up this professor in Aarhus and asked, “Can you give us a beam of phosphorus on it?” The professor said, “Yes, yes. No problem. It is going to come down the pipe at a few hundred kilovolts. We will give you a beam of phosphorus on it.” Meyer is on two phones and he said, “Okay. We’ve have it.” I said, “Fine. Maybe we can implant phosphorus into silicon and see if we can make an N-type layer out of it.”

So all of a sudden, we have this project. Meyer went back to parliament and said, “Yes. There is this American who says we can make PN junctions.” I had never said that. I just said that we would see what we could do. We did build PN junctions. It turned out that other people had done it before; I just didn’t know anything about it. A lot of what I have done is original but not new. But we got into ion implantation that way.

When I returned to Stanford, I assembled a group of graduate students and, like we would have done at Tungstal, we built a machine and we started implanting ions. But we did it the way that a device group would do it. We built junctions and we measured their properties. We had a strong electrical engineering circuit theory/device theory bias behind all of what we were doing, rather than the bias that a physicist would bring to it.

Everything we did was in IEEE. It was EE stuff we were doing, trying to build devices and then put them into circuits, whereas the physical review and the materials community had a different kind of people in it. When we began doing this, I started to receive invitations to go the nuclear physics centers around the world, being the only person who had ever figured out what to do with this, other than study materials with it. I would show up at these conferences, and all the materials people and the nuclear physicists would be there. I am comedy relief for this, last session sort of, “Here is what you might do with it.” I would come and talk about different devices and so on. We began to get the physics community into this business. It was an interesting way for people who otherwise only knew how to make bombs to think about something else. So they were thrilled.

Morton:

That’s a nice story.

Implications of new technologies for manufacturing; SUPREME (Stanford University Processing Relations)

Gibbons:

We got into ion implantation indirectly by going to a community that was trying to use it for energy, and we made semiconductors from it.

I was a consultant to these same eight guys then at Fairchild. I had students all over the place, and long after this was done, Gordon Moore, no less, at Fairchild and then at Intel when I was trying to get them interested in this, said, “Why would we every want to put a machine like that in a production line, with all the x-ray hazards and such, when diffusion technology that is more than accurate enough?” The beauty of ion implantation is that it gives you a dimensional control that is an order of magnitude better than diffusion, at least, and it is shallow. When junctions are one micron deep, the fact that you can make them a thousand angstroms deep and control them does not mean anything. But when all the device actions are occurring in the first thousand angstroms, then it means a heck of a lot. So we were way in advance of anything that anybody needed.

As is typical for all kinds of technologies, when a technology comes along that is a competitor for something else, nothing much happens as long as you can do it this other way, until all of a sudden a problem comes along that this technology will solve and the other one will not. That problem happened to be threshold shifting and MOS memories. If you want to have a system that is working at five volts, the threshold of the memory has to be two and a half volts, or something like that. It can not sometimes be four and sometimes one, which is what they were getting. You could adjust it with ion implantation. After you had made the device, it was the only way to do it.

Morton:

Was that something that you actually foresaw?

Gibbons:

No. We didn’t even do it. It was done by a man at Sprague Electric in Adams, Massachusetts. Bob Palmer, then working for Mostek , saw it and said that was the answer to the threshold shifting problem. Palmer was trying to build memory. He did it. Then there was a need to have ion implantation equipment in the fab. Then everything that we had been doing over the previous six or seven years was instantly available to everybody, and then all of a sudden this process just took off. Today, probably every single doping introduction step in a chip is done with implantation. Of course, by then we had range tables. We had done all the things that everyone needed to have to figure out how to design devices and how to build them. There were a lot of processor simulator programs. SUPREME is the general name for them, Stanford University Processing Relations.

We had all the computer tools you needed to be able to design devices and to choose implantation parameters that would produce certain kinds of results and go out and know that it was going to do that. It is a very quantitative technology compared to anything else. You count every ion when it goes in—you know exactly what you have, you know exactly where it is. It is very well done. It now has all the principles of nuclear physics behind it, and the ability to calculate things as accurately as you want. All of a sudden it becomes a futuristic technology. We would still be out there on the margins, except for the fact that we found something that could only be done in this way. It happens all the time in the introduction of new technologies. Less so these days because people expect to make a lot of money from new technology. But then you’re getting into the manufacturing area. You have to be able to make this work reliability every time do the same thing. That’s what is so good about it. Once you get over the problems of having these voltages around, and people learn that they work with it, it is okay.