Oral History Interview with Harold Kaufman (HK)
Conducted by Ric Shimshock (RS) on April 27, 2004

Harold Kaufman (left) and Ric Shimshock (right) at the
47th Annual Society of Vacuum Coaters Technical Conference

RS: Here we are at the 47th Annual Society of Vacuum Coaters Technical Conference in Dallas, Texas, and we have the honor of being with Dr. Harold Kaufman, one of the pioneers in the field of ion beam-assisted tools. We were hoping to get a feel from him on some of the contributions that this major development in the field has helped the rest of us get our jobs done.

Dr. Kaufman, how did you did get involved in the field? Were you trained in physics or some other area?

HK: I was trained in electrical engineering in WW II through an electronic technician program in the Navy, which was really first class. They left out calculus, but it was the equivalent of an electrical engineering degree with the latest technology. Then after the war I thought, “Gee, I think I’d like mechanical engineering better, so I took mechanical engineering on the GI Bill of Rights. Years later, I figured out that I had the equivalent of a double major.

I had no idea, of course, of going into vacuum coating or ion sources or anything like that. I started out in my first job after college with NACA, the predecessor of NASA. I worked on turbo jet engines for several years (afterburners, Reynolds number effects), which I found quite interesting. Then we became NASA, and I transferred into a new group that was studying electric space propulsion—how to convert electrical energy into thrust. I was given the job of taking a Von Ardenne source and seeing if it would make a thruster. The more I looked at it, the more I felt that the proportions were all wrong and it had to be done differently to be a good thruster. The Von Ardenne source was just too high on exhaust velocity for a good thruster. So I started working on what later became the electron bombardment source associated with my name. That was in 1958 or 1959.

RS: How did you acquire this Von Ardenne source? Was it electron based?

HK: It worked on the same principles as the electron bombardment source I worked on, but it was a single-aperture source. If you know anything about accelerating charged particles, you know you get very low currents out of a single aperture. But if you want enough current to be suitable for thrust to propel a spacecraft, you need less voltage and more current. You need to match it with the requirement. The electrons were emitted from a hot filament; they bombarded neutral atoms or molecules; the electrons were constrained by a magnetic field; and then they were electrostatically accelerated. All those parts were the same—just all the wrong proportions.

What I was able to do was to generate a large beam that could give significant thrust. That was in 1959 or 1960. Along about 1970, people became interested in modifying films with ion beams, so I had a head start on designing the sources that were needed for that sort of application.

RS: How big were some of the early thrusters and what were the dimensions?

HK: Oh typically, they were 10-cm beam diameter. We were interested in making something significant. We weren’t interested in small devices. There were, of course, some problems with the design group that I gave the job to. I told them the maximum pressure inside the chamber would be 10-3 mm of mercury (we used that measurement then, of course, instead of Torr). They came back with 1/8-inch stainless steel walls on the cylindrical parts and ¼-inch parts on the flanges. I decided that I’d better design it myself because the thrust-to-weight ratio wasn’t very good.

RS: Did you develop special power supplies for these devices too, or just the actual accelerating grid and the gas injection.

HK: The ion source and the combined gas injection/ionization region acceleration, these were the things I concentrated on. Actually, this was still back when we used vacuum tubes in power supplies. There was uncertainty about what we would use as a power supply because nothing looked suitable at that time.

RS: The original invention focused on space travel. Did you ever get to try any of these on actual spacecraft?

HK: There was a test flight in 1964, and there was another in 1970. But this is one of those fields where things move slowly. The first application did not occur until 1997. So, you can see that there was some slow motion there. I was involved in 10 or 20 mission calculations and program proposals and such. That’s one of the reasons I got interested in thin films because it was something I could do right away instead of waiting for 10 or 20 years.

RS: The time constant was much faster then. How did you transition from the NASA thruster approach to this ion source? Did people approach you or did you think that you might have something that could be applicable to the field?

HK: Well, a couple of people approached me. Jerry Cuomo thought it would be interesting to use these ion sources on some of the work he was doing at IBM Yorktown. He invited me to come out there for several weeks each summer. By then, I had retired from NASA, and I was working at Colorado State University.

For the first 8 or 10 weeks, I averaged one patent disclosure per week of employment. I got my post-graduate education in patents from Jerry. He was quite good at that. He taught me one of the fundamental things: If something goes horribly bad in the laboratory, you should say, “Now what can that be a real opportunity for?” You turn it around and look for a patent.

RS: Prudent advice. As you worked with these early films, what was the application that you were trying to engineer or accomplish?

HK: The early films were all really things that Jerry had been interested in. I was not a prime mover in any of the films. I worked with Jerry and with Jim Harper also, and later Steve Rossnagel—all at IBM Yorktown. Typically, they would present the problem, and we’d work together solving it. I sort of approached the films in a general way. I would come into contact with a lot of problems, and I would try to generate a few general concepts to deal with them, rather than looking at specific ones only.

RS: Did IBM allow you to practice your art in other areas or was it exclusively arranged?

HK: No. It was not exclusive. It was very generous in that we would have a patent, and IBM and I would have equal rights to the patent, which was very nice. This would help build up the patent portfolio. Very generous, and it’s easy to be productive under those conditions.

RS: Strong motivation.

HK: Yes. If you come up with an idea and you then sign away any right to ever use that idea again, that really does depress your creativity.

RS: What were some of the fairly early implementations of this ion source as applied to a vacuum chamber? Was it an auxiliary plasma that was rotationally symmetric?

HK: I think the first applications were etching, where you would be removing material. Typically, if you couldn’t etch it with a chemical (if it was something chemically inert like an aluminum oxide), you can still etch it with an ion beam. The difficult etching jobs were something where this was useful. Also, you could go to finer dimensions. I think Jerry Cuomo demonstrated a line width of like 100 atoms wide using ion beams back in 1976 or something like that.

High resolution was another option. Deposition was probably the next thing. You have a target; you sputter from it; and you’re not limited with any of the thermal constraints that you are with most other applications. It would be ridiculous but you could have a lead target and a tungsten target and deposit a mixture, atom by atom, of both of those. The lead would vaporize before the tungsten would melt if you tried to make an alloy of it. So, it gives great freedom in deposition.

The later applications tended to be improving properties. There is a general principle that one electron volt is roughly equivalent to 11,600 K. So, if you talk about 100 eV ion, you’re talking about something on the order of 1,000,000 K. So it opens access to reactions that you couldn’t possibly hope to see in an ordinary laboratory. I think that’s been investigated certainly with plasma processing. We did it quite early with ion beams. You can clean off a surface. There’s always a layer of adsorbed material. Just a short exposure to an ion beam makes much better adherence. You can compress a layer; bombard it as you are depositing it—this gives much greater hardness and improves the refractive index.

In some cases Jerry Cuomo was able to get metastable states that had not been observed before, and intermediate states in oxidation, for example, that were metastable and made a uniform layer. But you just couldn’t approach that with ordinary chemistry or an ordinary environment. So it opened up a whole range of new applications.

RS: Did you have to modify the device that you had developed for NASA for the different gasses or beam divergence? Was there a different set of design principles?

HK: The basic technology was the same, at least originally. You simply optimized it in a different way. I found it very stimulating to go back and forth from an electric propulsion to a thin film application. It forced me to look at things differently, constantly re-examine what I was doing and why. But as things have developed, the more recent trend has been (at least in ion-assist) to go to lower and lower energies. Those lower energies weren’t possible with the original ion source with grids. At present, I’m working entirely with gridless ion sources where you accelerate the ions in a neutral plasma. You are able to get much lower energies and still get useful quantities of current. With a gridded source, the space charge limitation drops the current drastically when you go to lower voltages, lower energies. So now I’m working mostly with gridless sources.

RS: Does that mean that there is a difference in the materials and the construction of the unit?

HK: Oh yes. With the electrostatic, you typically have two very thin grids, closely spaced with a lot of apertures in them, so you’re accelerating a lot of little beamlets in parallel. With the gridless type, you have electromagnetic acceleration where the acceleration results from the interaction of an electric current with a magnetic field. That establishes the electric field, and then you actually accelerate the ions in a neutral plasma, so there’s no space-charge limit. A relatively small source can generate, say an ampere of ion beam at an energy of 100 eV—something that you just couldn’t possibly do with grids.

RS: Well, I’m sure that who people have had to deal with grids that have burned out over time will be glad to look at some of these newer implementations.

HK: That’s another of the advantages. The grids were always a source of maintenance problems and breakdowns and failures. The gridless sources, if they will do the job (there are shortcomings – they don’t direct the beam quite as accurately as the grids) they are a lot friendlier to use in production. They sit there inert, and they have rather simple designs. They don’t fail often and it doesn’t require as much skill to maintain them.

RS: You mentioned that one of the early implementations was a Von Ardenne source. Are there other groups around the world working with some of these newer ion sources?

HK: There are a number of groups working around the world. The gridded source has long been in general use for a lot of groups. The initial work in gridless was done in Russia during the height of the Cold War in the 1970s. They were also electric space propulsion motivated. They developed closed drift sources. It was almost independent of our work. They did not develop gridded sources. We developed the gridded ones. Then at the end of 1970s, when we started to communicate again, we found that these two sharply different technologies had developed in the two countries.

RS: Are you aware if any of those have been implemented?

HK: They flew early. It was one of these things where they had a problem, and they solved it. We didn’t have the problem, so we didn’t solve it. We used chemical propulsion, and we had a well-funded program (the hydrogen-oxygen, the high-oxygen alcohol hydrocarbons, etc.). We also developed storable propellants. Putting a satellite in orbit, we were able to get good propulsion with something that was sitting up there for a long time.

The Russians did not get storable propellants, as near as we could figure out. We’ve tried to figure out the sequence of events afterwards. They ended up using ion thrusters of the closed-drift type in practical applications in about 1964, about the time we were making the first flights with the gridded sources. By now they have flown (a couple years ago) over 100 of these sources in satellites. But their choice was, they didn’t have storable propellants. They either had to use compressed gases, which are terribly inefficient for propulsion, or they had to go to electric thrusters. They chose to go to the electric thrusters. We had the good storable propellants; we didn’t use electric thrusters.

RS: So it was tough to compete with that technology from your early implementations then?

HK: Well, it’s a different exhaust velocity range. In space propulsion, you want a certain specific impulse for a certain mission. We actually were focused primarily on interplanetary missions. For interplanetary missions, the sort of exhaust velocities obtained with gridded sources still makes sense. However, the Russians (I use “Russians”; it was actually a Russian development pretty much, almost all of the people involved in it, even though it was the Soviet Union back then) The Russians focused on lower exhaust velocities—somewhere around 1000 to 1500 seconds. We used specific impulse. That’s more suitable for satellites. So what they did was actually focus more toward satellites, which we were kind of ignoring back when we were originally working on it. Just like we ignored the power supplies and had vacuum tubes.

RS: What do you think is one of the major accomplishments as you look around the industry and see where all of your sources are implemented? What makes you proud of some of the different devices you have developed?

HK: Oh that’s a difficult question. You must realize that a lot of us work in technology because we enjoy it. It’s a nice way to make a living. We get caught up in the pursuit. I’m not a hunter, but I suspect that looking for new knowledge must have some of the same thrill as going out and hunting for a wild animal. You get caught up in that and you don’t worry too much about some of the other things. I try to take an overall look. What I get out of the overall look is that in technology, we are a subculture of the overall whole culture we live in.

A couple of years ago I visited Mesa Verde. That made a deep impression on me. There were stone houses. People lived there in relatively secluded areas and were able to sustain their culture from about 600 A.D. to 1200 A.D. Except for little changes, like maybe some human sacrifice at the end, the culture was essentially static. 600 years. It only took us 300 years to go from Sir Isaac Newton to landing on the Moon. Something has changed in the last few hundred years so that we’ve become a very active society, technically moving ahead. Still the people who are doing it are a small fraction, maybe 1% of the total culture. I guess what impresses me more is not my particular contribution to it but rather that I am part of a subculture that is moving ahead our overall culture at a terribly fast rate compared to anything previous in history.

RS: As you mentioned the future, where do you think that this technology will go? Do you foresee more use for interplanetary emissions for these thrusters? Do you see ion-assist low-voltage new devices for physical control of nanostructures?

HK: I think both of those. We are planning at present for interplanetary missions. We’re looking at all types of thrusters. That seems to be moving ahead. That one is surprising how slow it is. We landed on the Moon 35 years ago, and we never went back to it. But we do seem to be moving ahead on some of the plans to go to other planets.

I see partly the end product changing. We’re obviously going to have “smart” machines in almost everything we touch. There will be computers embedded in everything. The ion sources will increase their use. They’ve generally been competing with plasma processing. If you can do it with plasma processing, it’s probably cheaper than having a separate ion source. But the ion sources are coming down in price; they’re becoming more useful and more practical. I think that they will continue to expand in use. So will a lot of other things.

In the ion source field, it seems like we add new ion sources, new concepts, all of the time; but we don’t drop many of the old ones. The whole field becomes more complex.

RS: And broader. We’re going to conclude at this point, so if there is anything else that you would like to mention for posterity . . .

HK: Well, I’ve enjoyed this meeting, and I have particularly enjoyed the tutorial seminars. Don Mattox looking back on some of the technical developments was very interesting. Charlie Bishop’s talk was fascinating too. He’s looking at transition. Our profit model for the past is going to disappear. We’re going to have to live more as a commodity rather than high-tech where we can charge very high prices. That’s part of a change that’s occurring. We’ll keep seeing more of those changes too. We seem to be becoming more competitive and the rate of technological development doesn’t seem to be slowing at all yet.

RS: Thank you, Dr. Kaufman. I appreciate all the work you’ve done to help make our jobs easier in the field and make these films that people are asking us to fabricate.