AVS Historical Persons | Eric Kay - 2007

Eric Kay - 2007

Oral History Interview with Eric Kay

Interviewed by Paul Holloway, October 16, 2007
HOLLOWAY: Good afternoon. I'm Paul Holloway from the AVS History Committee. Today is Tuesday, October the 16th, 2007. I have the pleasure of interviewing Eric Kay, retired from IBM San Jose and more recently from Stanford University. We're at the 54th AVS International Symposium in Seattle, Washington. So Eric, welcome, and thank you very much for agreeing to the interview. 

KAY: Thank you.

HOLLOWAY: I was wondering if you could start off by giving us some of your background and occupational history.

kay.jpgKAY: Okay. I got my PhD in physical chemistry, specifically, in high-temperature thermodynamics, and when I graduated, I thought I had arrived. Since I wanted to stay on the West Coast, I applied in some of the more chemically oriented places on the West Coast, which at that time were mainly places like Shell Development and Standard Oil and that kind of place. But at Shell Development, the job for which they wanted to hire a thermodynamicist was, quote: "someone who can help us figure out why wax cracks on milk cartons." That was very disillusioning. First of all, I couldn't figure out what high-temperature thermodynamics would have to do with this problem, and secondly, I had great dreams about doing something more consequential than that. 

HOLLOWAY: You didn't think milk cartons were your future, then?

KAY: No, Anyway, even though Shell Development had a very good reputation for supporting basic research. I somehow hoped I would be given a more global challenge. So fortunately, I heard about formation of a new research branch for the IBM Corporation, on the West Coast in San Jose, and so I interviewed there, and was incredibly impressed. It was in a very contemporary sort of structure with lots of paintings and sculpture all over the place, that is, a very appealing setting in which to work. The man who interviewed me said, "Hey, we do not have any people with chemical background. We really need to fill that gap. And when I asked him, "What kinds of things do you have in mind for me to do?" he told me - as true as I'm sitting here - he said, "I haven't got the vaguest idea. We want you to tell us what a chemist can do in our laboratory that might be of eventual relevance to IBM." 

HOLLOWAY: That's a remarkable beginning.

KAY: Yes, and I thought that was a far more attractive challenge than 'cracks in milk cartons.' 

HOLLOWAY: So you were one of a few people in that organization.

KAY: Yeah, at that time, there was already an applied IBM lab in San Jose, but in the newly formed research lab, which was the new West Coast arm of the IBM Research Laboratory on the East Coast, there were only about six, or seven of us. I was number seven to join that staff. One was an MD; one a psychologist; one a linguist; one a mathematician; one a solid state physicist and the guy who interviewed me was a theoretician, physicist. I was told, the main mission in those early days for this newly formed research lab was to establish a reputation for IBM on the West Coast in sciences or related fields in order to attract some of the graduates from places like Berkeley and Stanford and Caltech, most of whom in those days - 1958 - were inclined to say, "I am not moving away from the West Coast". Anyway, that's how I got to IBM, and since I was totally at liberty to identify what I might want to do, I nosed around as to what kinds of things IBM was doing out there, to see in what way I might relate to that. At that time, early1958, IBM also had a manufacturing plant in San Jose which, together with the people in their applied lab, put IBM on the map by pioneering the hard disc magnetic recording industry. 

HOLLOWAY: Now, this was tape recording or disk recording?

KAY: Disk recording, Since I came with a chemical background, that naturally drew my attention to the magnetic storage media they were using. They were primarily interested in the magnetic characteristics of these materials and, of course, the very demanding mechanical and electrical issues relating to the machinery in these recording devices. Surprisingly, they seemed to have only very superficial insight into how chemical materials characteristics related to the desired magnetic properties. They bought the magnetic material for the storage medium from an outside vendor, and then suspended it in a liquid slurry which was then applied to their large hard disc substrates by a centrifugal spinning process to end up with a magnetic layer which was many microns thick. The magnetic component of this slurry was gamma Fe203, which is an allotropic form of the much more common alpha Fe203. .Allotropes are chemically identical but crystallographically quite different. Gamma Fe203 has a spinel structure, and only it had some of the magnetic characteristics needed. Alpha Fe203, has a hexagonal structure and is antiferromagnetic, and therefore of no interest in this context. I had the good fortune that in the applied lab nearby, there were a number of physicists who, in contrast to me, understood magnetics pretty well. I very quickly learned that in the future, if more high-density storage was to be accomplished, one would have to make the magnetic medium very much thinner than they could possibly attain by their spin coating process. So I might say one lesson, in hindsight is, if you ask enough questions and familiarize yourself with what the future key needs are, that often, helps you decide how you might impact it - not this week, not this year, but maybe in years hence. 

So, I decided that I would look into how can one make a much thinner layer in a controlled compositional and structural fashion? In those days, 'thin' meant several thousand angstroms. Of course, nowadays, one has to be able to do all of this in a magnetic active layer that's maybe a 100 angstroms thick, or less. But that was in '58'. Obviously, some sort of vapor deposition process suggested itself, and so I started reading the literature. I really knew nothing about thin-film anything. I mean, that was not a high-temperature thermodynamics specialty. Eventually, I came across this German paper of somebody using plasma discharges to make compound thin films, iron oxide in particular, albeit the wrong oxide. I should remember the guy's name, because he's sort of the father of film deposition in plasmas.

HOLLOWAY: Fred Wehner.

KAY: No, no, long before Fred. This guy did it in the late 1800s. I think the guys name was Koenig. In German, the process was called "Kathodenzerstaeubung" ie.' sputtering' in English - a lousy translation. I tried to change that when I was invited in 1962 to write a book chapter in the annual edition of "Advances in Electronics and Electron Physics" about "sputtering" and I titled my chapter "Impact Evaporation and Thin Film Growth in a Glow Discharge". I don't think many thin film people read it and obviously the term 'impact evaporation' didn't catch on.

Anyway, back to 1958, to a 'high temperature' thermodynamics guy like me, chemistry in plasmas sounded pretty intriguing. Whereas, at that time in the AVS and also at IBM East Coast there was a gang of people like Klaus Berndt and Neugebauer and Francombe and others, talking about magnetic thin films made by thermal evaporation, nobody seemed to know anything about it in a plasma environment, so that made it even more interesting, and this Koenig publication claimed to be able to make compound thin films with controlled structure and composition. So, in 1958, I was certainly the first person to introduce plasma sputtering to IBM, West Coast, and as best as I know, to IBM in general. So, that's how I got into sputtered thin films.

HOLLOWAY: You were depositing iron oxide?

KAY: I was trying to make gamma iron oxide, and as I mentioned earlier, making alpha Fe203 is trivial - I mean, there is a huge parameter space within which you can make αFe203. But the trick was to make this totally different crystallographic form of the same chemical, namely, gamma iron oxide, since it has a lot of unique magnetics advantages. Fe304, which also has a spinel structure and magnetically interesting is less desirable than gamma iron oxide, which has to do with magnetic shape anisotropy of the individual crystallites. But I don't want to get too detailed with this. 

So anyway, so I built up a thin film facility, and I was very fortunate that the guy that hired me in 1958 was incredibly bright and open minded and generous, because within a few months, I convinced him that besides the thin film plasma system, I needed a X-ray diffraction and fluorescence facility, an electron microscope, a mass spectrometer, a quartz microbalance, and several things of that order, none of which I had ever used before. The thing that just flabbergasted me, after only six months at IBM and after a full day's presentation to upper management I was told "Write it up, and we'll get it," whereas, in my PhD days in graduate school, if it cost more than $10, my Professor would more than likely say, "You can't have it." So this was a whole new world. 

HOLLOWAY: It would be a whole new world versus what you can do today as well.

KAY: Exactly. it started me out with this naive impression - and that's why I'm bringing it up - that money was no object. One incident, which may be of interest historically of what can happen if you're exceptionally naive, as I was at that time. In those days, nobody really knew what was going on in any detail, either in the plasma gas phase, or the gas phase interacting with a surface. I mean, you could do some name-dropping, but there was no real science available yet. So it was kind of starting from scratch, as far as I was concerned. But nevertheless, since I, on occasion, succeeded in making the right stuff with the right magnetic properties, I had the audacity to actually go to the applied guys, and gave them this big, profound physics rationale as to why eventually they're going to have to stop making their magnetic medium by the process that they used and switch over to a way of making much thinner layers - e.g., by some sort of a 'vacuum' deposition process. They actually showed interest in my results, but when I asked, "Well, when will you be ready do something like this?" they laughed and almost threw me out of their office, because not only did I have the audacity to tell them that I allegedly had something better than they had, but also I obviously had no idea then , what it involves to switch technologies. One of the guys took me out onto the plant site where they made these large magnetic recording disks, and here was a row as far as the eye could see of machines that made these things by this liquid centrifugal spinning process. Anyway, he said to me something like, "Now, do you really think we're going to shut down and throw away this many million dollar investment, and that we're going to go to 'vacuum' systems in a manufacturing environment?" I mean, in 1960, they thought I was crazy - not that I hadn't done a nice materials science job, but for me to think that just because I can make a much thinner layer of the right stuff on a small glass substrate in a fancy vacuum chamber would be enough for them to make major changes, was naive, to say the least. 

However, I stayed with that general approach of thin film synthesis in sputtering environments. By now, I was actually encouraged to hire a couple of other people who could work with me. I remember , besides the plasma approach, around 1966, Erich Sawatzky and I, published some optical emission spectroscopy we did on sputtered species coming off various targets, using a fancy duoplasmatron ion beam for sputtering a target held at 10-7 Torr, an electron beam to electronically excite the sputtered species and look at them in a spectrometer. For us at least , that was quite an eye-opener. We sputtered metals, alloys and compounds. and then , of course, tried to correlate this data with what we saw ending up in a film made in a plasma environment. In 1973 Schoch and I published a related matrix isolation optical spectroscopy in a Triode plasma system. Making the correlation to film structure turned out to be less than straight forward which, eventually, led me and Gary Selwyn to do spatially resolved optical emission spectroscopy right in the plasma, using fiber optics to look at species all across the plasma as well as those arriving at a substrate in this much shorter mean free path plasma environment. 

Utilizing actinometry, which John Coburn and Martin Chen had introduced came in handy. 

That was another eye opener for us, especially in later years, when I got into plasma polymerization. 

Of course, I recognized early on, if you're going to make much thinner films, characterizing them in that form was going to be a key issue. That brought up the need for electron diffraction, electron microscopy, and X-ray diffraction and X-ray fluorescence, microbalance, to understand composition and structure on a microscopic scale. And so I became familiar with those sorts of things, and was also encouraged to hire some more people to help me.

HOLLOWAY: So all of this was in the early '60s?

KAY: Yes, I think somewhere in the '60s. And, since we're talking about magnetic materials in those days, we used the electron microscope not only to get morphology by microscopy and crystal structure by electron diffraction, but it was also the beginning of Lorentz microscopy, which allows you to look at microscopic magnetic structure. If you want to understand magnetic spin reversal processes involving moving domain walls, then Lorenz microscopy was really ideally suited to do that. So I got into Lorenz microscopy. But then, as often happens in industry, if you do a good job, they ask you to become a manager. Then they let you hire other people. I agreed to be a manager, provided I could keep my technician and could get a one year Post Doctoral Fellow on an ongoing basis to help me keep my own research going. But of course, when you are in charge of hiring other people, you have to figure out, what in general you want them to do. By this time, in the late 60's, the days at IBM of saying to an applicant, "I haven't got the vaguest idea as to what we want you to do, provided it was good science'' were over. So, I basically had to know what is needed in order to make this whole area grow to eventually build a competence not only in thin film science but in surface and interface science in general including the liquid solid interface. 

One of the next things that happened around that time was the realization of the importance of generating your materials in clean environments if you want to make a well-defined structure and composition in order to be able to control physical properties on a microscopic level - in this case, magnetic properties. Clean, in those days, meant background pressures of 10-7 Torr. I remember, already in 1965, Bill Gill and I published a paper on 'Efficient Low Pressure Sputtering Using a Large Cylindrical, Inverted Magnetron Suitable for Film Growth'. This DC sputtering configuration worked effectively down to 10-6 Torr before the plasma became unstable for space charge reversal reasons. Of course, later on, John Thornton's RF Magnetron configuration was much more versatile and practical and became the workhorse for everybody.

HOLLOWAY: Vacuum? 

KAY: Yes, vacuum. But as we started calculating with how many layers of background crap you may be bombarding a film's growing surface every second , we soon realized that probably 10-6 Torr wasn't clean enough. And it was about then, and I may have the dates a little wrong because my memory isn't that good anymore, but it was in that general time period where ultra-high vacuum was becoming available. You know, Varian Vac-Ion pumps and that kind of thing. 

HOLLOWAY: Vac-ion pump was invented, if I recall correctly, at the end of the '50s, and so the '60s were at the right time for the high vacuum becoming very standardized.

KAY: So then we recognized that if we really want to control thin-film properties, we've got to do it in a cleaner environment but also, of course, understand and gain control over many other aspects. Magnetics, and every other physical property - electrical, mechanical, optical etc, - they're all crucially dependent on controlling structure, and not just crystallographic micro structure. In magnetics especially, even the macroscopic shape and size of crystallites are a very important things to be able to control. And of course, the composition and stoichiometry and getting the cations in these oxides to end up in the right lattice sites and all of that kind of stuff. At that very same time people really began to realize that with these high 'surface to volume structures ', surface and interface physics phenomena, especially at higher temperatures , are very likely to play a significant role in the film's overall characteristics. Of course, nowadays, everybody acknowledges and knows that. But in those early days, just how different and how important a role these surface and interface properties would play was not well understood. Physicists, sort of knew there ought to be a dramatic change at these structural discontinuities, but I don't think we realized at that time how critical a role those differences are likely to be, especially in multilayered film assemblies. And, of course, consequently, we recognized, that UHV is critical as far as many interfacial phenomena are concerned, like for example in epitaxy. Also, prior to this time there weren't any tools yet, that allowed you to selectively look at the unique characteristics of a surface. With the advent of ultra-high vacuum, all the escape-depth-dependent electron spectroscopies, like photoemission, LEED and Auger and all of that stuff was evolving. Other people were developing these surface science characterization tools, all of which depended on ultra-high vacuum. 

And so, over time, we began to adopt these, and since we were looking at magnetic properties, I got involved not only with ESCA (electron spectroscopy for chemical bonding surface analysis) and Auger to do elemental compositional surface characterization, but also to do spin-polarized surface electron spectroscopy, because that allows you to uniquely look at the magnetic hysteresis characteristics at a surface. Surface, in those days, was defined by escape depth. This was before glancing-angle core electron spectroscopy, XPS, or at least we didn't have that at that time. So it was several surface lattice layers from whence your signal came. We were impressed and amazed of how different the surface-magnetic characteristics are, especially as a function of temperature, compared to the bulk magnetic properties of the film. For bulk thin film magnetic characteristics in those days, we used Kerr magneto-optics, where the light penetration is much deeper than the escape depth in spin polarized electron spectroscopy. So, by having all these techniques built into the same UHV vacuum system, we were able to compare surface phenomena from bulk phenomena in situ in the presence of ion bombardment during film growth, and that, we felt, was pretty exciting in those days. I actually continued to do this kind of study with my graduate students after I retired from IBM in 1991 and joined the Stanford Faculty. IBM let me take all my research gear with me to Stanford.

For me in the 70's another question was -- since we now have fairly clean environments and I had good luck with growing sputtered 'epitaxial' garnet films on single crystal garnet substrates at considerably lower temperatures than was the case by evaporation, can one take advantage using energetic particle bombardment in a controlled fashion, to induce low-temperature epitaxy on an amorphous substrate? People were always telling you, "Boy, if one could make this stuff at room temperature in single-crystal form on glass that would be fantastic." And you know, it didn't take much imagination, if you knew a little bit about film nucleation and growth, that you should be able to influence atom surface diffusion during condensation and a whole lot of other things, by energetic particle bombardment during film growth, but to what degree.?. And so with my PostDoc, Paul Ziemann, we very much got into how can we demonstrate what ion bombardment can do in terms of controlling film growth mechanisms during film growth? We did ion beam experiments as well as trying to duplicate results in a long mean free path plasmas by biasing the substrate in a low pressure triode plasma system to approximate monoenergetic ion bombardment. But, of course, in a plasma, there are always so many more things going on simultaneously that the best you could do was to phenomenologically demonstrate what happens. In these plasmas we were able to grow highly oriented films at room temperature on glass with indistinguishable mosaic structure from bulk single crystals, but only in a direction perpendicular to the plane of the film, with much more random structure in the plane of the film. During this work, we also learned a lot about point and line defects induced or alternately annealed out by ion bombardment depending on the energy and flux of the energetic gas ions per condensing particle during film growth. We saw similar results for unit cell lattice parameter distortions in the films which, of course, reflected itself in many physical properties, for example, in the superconducting transition temperatures of metals. However, all told, the semiconductor physics people who were the 'epitaxy' fanatics at that time, weren't too impressed. 

This experience was reminiscent of some earlier work,1969, when Don Raimondi and I had published a paper showing that one could 'fine adjust' stoichiometry of oxides in an oxygen plasma by stuffing more oxygen atoms into the lattice with controlled ion bombardment, resulting in all kinds of changes in physical properties of the resultant film. For example, increasing the dark resistivity of zinc oxide, which is normally oxygen deficient, by many orders of magnitude. However, the fact that the accompanying induced defects partially killed the photoconductity, made the whole venture uninteresting to our IBM's electrophotography community, whose needs we were trying to impact. So we didn't pursue that any further either.

I mean, it's sort of interesting to think back when these things were novel , because, of course, nowadays, if you told this sort of thing to one of the guys who's in this business, he would say "Well, what else is new?" 

HOLLOWAY: Yeah, self-evident.

KAY: Right. But for us it was very exciting. We weren't the only people that ran into all this surface stuff - of course not - but we made use of it as soon as we became aware of it, and that just opened up whole new dimensions for us. And then, of course, since my thin film synthesis interests very much included thin film growth in plasmas, I had the great wisdom to hire John Coburn and Harold Winters, and although they had totally different backgrounds, just as I did originally, I did talk them into joining my group. I told them that I was the local thin film nucleation and growth plasma guy, but I needed people who can help me understand better what goes on in the vapor phase of the plasma, that is, do thorough plasma diagnostics and help me understand how all that related to controlling film growth in such an environment. John Coburn with electrical engineering background liked that idea, and he, as most people by now know, has really become an expert in this plasma diagnostics area. Harold was more of a surface science UHV oriented person. I think I can take credit for guiding Harold to work with monoenergetic ion beams in an ultra-high vacuum environment and to try and understand to what degree especially surface sorption phenomena may differ when you superimpose an energetic bombarding specie on top of the conventional condensation and growth processes. This helped us understand gas trapping as well as a few aspects of reactively sputtered films, earlier than many people. Harold, like John was very talented and built a very sophisticated UHV ion beam facility to study energy and current-density and directionally controlled beam/surface interaction processes. John and Harold together certainly, I think, were pioneers in plasma/surface related interactions in those early days. And so the three of us looked into many of the things that we felt you needed to know if you really want to know what goes on during film growth in a plasma environment and we published quite bit together. I don't remember the timing now, but somewhere in there the whole business of sputtering, reactive sputtering etc. caught on in many other places including all the variations of DC and RF and microwave plasmas, all of which were evolving at that time, and all of them needed plasma diagnostics desperately. Langmuir probes, turned out not to be terribly helpful, because, especially in a reactive environment, they were difficult to use, so one had to figure out other ways of measuring ion currents and energies and electron densities and also be able to identify chemical species arriving at the film growing surface held in the plasma. John's electrostatic ion energy analysis and mass spectroscopy attached to the plasma chamber did wonders for us as a diagnostics approach. In this general plasma related diagnostics area it was where John's diagnostics work and Harold's beam approaches set the pace.

Gary Selwyn and I were also among the early users of spatially resolved optical emission spectroscopy with fiber optics to do spatially resolved chemical monitoring across plasmas. 

HOLLOWAY: How big was your group?

KAY: At that time, that is in the late 60s, I actually had added a number of other disciplines to the group since I was trying to build up a much broader Surface Science capability besides Thin Film and Plasma Science. At that time it was probably like a dozen PhDs., but only Harold and John and I were working on the plasma and related thin-film aspects. Eventually, we had other experimental surface scientists, like Dick Brundle and Dan Auerbach, and also some solid-state physicists who were looking at various surface sorption and dynamics and solid state physics phenomena. I also managed quite a few theorists, who among other things liked the easy access to the latest computational facilities at IBM.

Eventually I guess, by middle 1980s, my group had something like 50 PhDs in it and quite a few one year Post Doctorial Fellows from all over the world who brought in specialty knowhow. I put several of the theorists into project management positions to help me encourage scientific rigor and at the same time make them more aware of the real world. By now we were also into surface/liquid interface studies as encountered in electrochemistry and also started an effort in tribology. 

But Paul, I mean, is all this stuff too detailed, to discuss right here?

HOLLOWAY: Go ahead. So was there a big community outside of IBM innovation that was working on the plasma area?

KAY: At that time -- oh yes. I think glow discharge deposition - sputtering -  there were a lot of people getting into it, both in IBM and in all kinds of other places. But within IBM, which is the environment that I knew best, I think we were the people trying to understand to the best of our ability what goes on in this incredibly complex plasma environment. Now, there were other people at IBM, for example in Fishkill where IBM had a huge manufacturing facility, who were using plasma processes for technological purposes. You know, people like Maissel, who was certainly one of the sputtering pioneers. But they had to deal with a whole different set of issues about which we knew nothing, that is, how do you translate what you can accomplish in a small, very well-defined piece of apparatus with all kinds of in-situ diagnostics techniques into a manufacturing environment where a single mistake can cost you millions of dollars. I mean, in those early days depending on sophisticated in situ spectroscopy diagnostics, for example, was unthinkable in a manufacturing environment. Now, it isn't unthinkable. Now, it's unthinkable to do otherwise. I have sort of a nice incidence on that, trying to talk the manufacturing people into using optical emission spectroscopy in all their RIE systems, which is kind of funny, but I'm getting a little ahead of myself. 

I think it was in the late seventies or so , when the semiconductor microcircuitry world recognized, more and more, that liquid etching for making microstructures just wasn't adequate for all kinds of reasons, but let's say, primarily for resolution reasons, and so the idea of trying to do dry etching began to be an important alternate concept. And of course, then eventually that led to 'reactive ion etching' RIE, and guys like Harold Winters and John Coburn, were some of the key people in those early days, to provide some insight of what mechanistically goes on in these enhanced surface chemical etching processes in the presence of energetic particle bombardment.

HOLLOWAY: They certainly have been acknowledged as being the people who led that field. 

KAY: Absolutely. 

HOLLOWAY: I know that John and Harold both acknowledged that you were a partner.

KAY: Well, my main contribution by that time was to recognize that what they could do would be very important and my job was to create and sustain an environment so that they could go ahead and do it. But I take very little credit for what Harold and John then went into the lab and did relating to RIE. I mean, we obviously discussed it all the time. I was their manager and all of that. But on the other hand, they were clearly the guys that put that on the map, and I think to this day, their work stands the test of time. Obviously, by now many other people have made important contribution to this whole area, all over the world. 

HOLLOWAY: Absolutely.

KAY: Anyway, by this time, as I said earlier, the semiconductor microcircuitry world was getting into trouble with how to remove, that is , etch materials with high spatial resolution. Prior to that, the emphasis in vacuum approaches was mainly on condensation and growing thin films by molecular beam epitaxy and doping issues. To this point much of our own work was also primarily motivated by thin film growth issues, but in plasma environments. And then, of course, came reactive ion etching, RIE, with its ability to dramatically change rates of chemical etching reactions beyond physical sputtering, depending on such things like the energy and direction and flux of the incident ions. If I remember rightly, this enhancement was first demonstrated in fluorocarbon plasmas by several people. We first published on it in 1975. Soon, everybody was trying to understand , first and foremost how fluorine in plasmas interacted with silicon -  beyond the simple minded concept of Si + 4F makes SiF4 and then also, of course, what is happening to the carbon species. And it's interesting that to this day, people are still trying to understand the mechanistic details of these pretty complicated surface reactions. And of course, by now people work with many other etching species on many other substrates, all of which have their own little special problems to resolve. 

HOLLOWAY: I recall when that came out, that there were sessions organized at the AVS meetings on this stuff 

KAY: Oh, absolutely. 

HOLLOWAY: Did that occur naturally, or were you active in the AVS and drove that? 

KAY: I was active in the AVS, but I don't think I could take much credit for promoting any sessions uniquely along those lines. I mean, I was an AVS Director for a while, and I guess I was in a number of program committees over the years. So in hindsight - you know, I've sort of forgotten, I guess I must have had something to do with arranging such things on program committees , you know, who to invite to talk and that kind of stuff - but I don't feel that I can take any great credit for any of that. I did my thing, as I said, as a Director, but that was on a whole set of problems - not just plasmas.

Since you are asking though –did I drive any of this stuff ? - I was the first person to promote and chair the first Gordon Research Conference on 'Organic Thin Film and Solid Surfaces' and as one of the 'Founding Fathers' of the Material Research Society, MRS, I guess I did contribute to alerting people, especially, to all sorts of thin film science issues.

By then, everybody in the surface science world was tuned to doing everything in ultra-high vacuum as part of the 'otherwise it wouldn't mean anything kind of attitude'. I can kind of remember when I first went to Harold Winters and reminded him, "Hey, Harold, you know, they use fluorocarbon as etchants --they'll never use fluorine gas as such because the safety people wouldn't let them--- and they'll never use xenon fluoride because it's too expensive, so couldn't you extend your studies to include carbon containing etchants, like CF4. Initially, he vehemently refused because he wasn't going to crap up his system with anything 'organic'. This was before it became obvious that the byproducts of these 'organic' etchants were going to play an absolutely crucial role in via etching, as protection layers.

HOLLOWAY: That's the side walls.

KAY: Exactly, but also enhancing selectivity in etching various materials. I'm just sort of trying to describe what the mentality was at that time in the surface science UHV community that I moved in. These were primarily surface physicists, not chemistry background catalysis people. I mean, the people in the surface science world that I knew at that time just thought I was crazy to suggest putting a carbon containing specie into their system. Highly corrosive xenon fluoride was OK?

HOLLOWAY: A lot of people still hold that attitude. They won't let organic materials get close to their ultra-high vacuum.

KAY: But it's changed. It's changed a lot. Just look at what the bio-interface guys, like Michael Grunze, in the AVS are doing these days. So anyway, where were we? So eventually we all ventured into the fluorocarbon world, for, I think, by now very well-known reasons. Oh, I know what I was going to say. John Coburn and I had already learned a lot by 1976, using his fabulous plasma diagnostics system, about what an incredible array of low molecular weight 'organic' fragments are produced in one of these etching plasmas. Since I was more of a chemistry background thin-film-generating guy than an etching person, I was wondering, if you use fluorocarbons for etching, you're very likely going to create gas phase precursors for polymerization among many other fragments in the plasma and ultimately make polymers somewhere in there. Other people had already made polymers in plasmas, but in totally different contexts. Initially, in the RIE world, polymerization was considered a pain in the neck, because who wants polymers in their etching system? But, of course, we and everybody soon learned otherwise. We learned that inevitably you will form some flurocarbon polymers. So the only thing that you can hope for was to minimize or optimize it, but you can't totally prevent it. But then, of course, emerged this whole idea "Hey, this has some good consequences, like side wall protection." And then came the realization, that both the fluorine etching process and the fluorocarbon deposition process go on simultaneously at all surfaces exposed to the plasma, and all you have to really learn is how to regulate and control the degree to which these two opposing processes took place at various surfaces held at different surface potentials with respect to the plasma potential. So with another Post Doc, I got quite involved with trying to understand a little bit more about the mechanism of plasma polymerization.

HOLLOWAY: Was any of that ever used in magnetic recording media? 

KAY: Yes by others- as protective tribological layers on top of the magnetic media. Hard carbon coatings, are made using either hydrocarbon or fluorocarbon plasmas with high-energy ion bombardment of the growing film to get rid of most of the fluorine or hydrogen and leave behind a dense, amorphous hard carbon film. But that wasn't my motivation for doing it. I just wanted to understand what controls the formation and structure of polymers made in the plasma, for whatever reason you wanted to use it. Over the years I published a lot of stuff on this plasma polymerization topic in various plasma modes. 

However, what eventually really tickled my fancy a lot, came from my realization that in a very conventional RF diode plasma sputtering system, you could sputter a metal or anything else off a target by conventional argon sputtering processes and simultaneously you could also create appropriate polymerization precursors in the plasma, if you added a partial pressure of the right hydrocarbon or fluorocarbon gas, to the argon. So, in principle, in an RF diode system you could sputter metal atoms off the target electrode at high incident ion energies and simultaneously produce a polymer film of your choice at the low ion energy, grounded substrate electrode, thereby allowing you to produce a metal containing polymer film. The trick was how to keep the metal target clean so you don't get the hard carbon coating permanently poisoning the metal target electrode. There are some obvious ways of utilizing RIE and sputtering to accomplish that, and growing a polymer film only at the substrate under various substrate biasing conditions to control the polymer structure and rate of formation. I started publishing this kind of stuff in 1976. For me this opened up the whole area of granular composite thin films, that is , a uniform dispersion of granular inorganic material A in a granular polymer matrix B - in film form. I guess we now refer to these granular particles as nanoparticles. In those days, we called them clusters. Since the metal arrives on the polymer film surface during film growth as atoms we had to understand the aggregation of these atoms into controllable-sized clusters inside a polymer matrix, involving Ostwald ripening-like processes. Picking a Teflon-like polymer matrix which does not readily form chemical bonds with the metal certainly helped with understanding these nanoparticle aggregation processes. And if you can learn to disperse metal particles with controlled size and shape in a chemically 'inert' dielectric film matrix, then there are all kinds of electrical, optical and magnetic physics phenomena that you can study, that depend both on the volume fraction and the size of metal clusters within the polymer. This ability allowed us to test a number of different so-called Effective Medium Theories intended to predict these subsequent physical properties. Independently, we also learned how the mechanical properties of the polymer film is modified by incorporating various volume fractions of an inorganic clusters into the matrix. There is this percolation phenomenon - I'm sure you're familiar with it - which occurs at a unique volume fraction at which, for example, electrical, optical, magnetic and mechanical properties of the composite film change dramatically as the metal clusters begin to communicate with one another in various ways. So, for example the electrical conductivity at the onset of percolation changes by many orders of magnitude as the electrical properties of the composite film are suddenly no longer dominated by the dielectric matrix but by the 'short circuiting' metal clusters.

After publishing many papers both on the synthesis and characterization and especially on the electrical and optical and magnetic properties of these granular composite metal/polymer films both above and below the onset of percolation, we began to ask what contemporary practical problems could one solve with these systems. The first one we worked on was 'circuit writing in air using scanning laser beams' or for finer line resolution, using electron beams in vacuum. We had learned a lot about migration of isolated small metal clusters within these dielectric polymeric matrices as a consequence of annealing them, such as their sudden coalescence to form larger aggregates of metal clusters above the glass transition temperature of the polymer. Using a laser beam as a spatially confined intense heating source allows you not only to cause this metal cluster aggregation but also the removal of the polymer by ablation from the locally heated area. So by starting with a film containing totally isolated metal clusters, that is, at a below the percolation volume fraction, where electrical properties were totally dominated by the dielectric polymer matrix, we could produce highly conducting local metal regions as a result of coalescence and at the same time get rid of the dielectric polymer in the heated region. Then we demonstrated that by scanning the laser across the film allows you to write a highly conducting metal line exhibiting bulk conductivity. 

My postdocs and I did a lot of work on this approach and we thought, we will surely find somebody in the applied community who would be interested to get involved, but alas, we pretty much ran into the same response "You've got to be out of your mind to really think we're going to even contemplate throwing away our established approach of writing circuits, for example, on circuit boards, just because you can write very much thinner conducting lines in air using a scanning laser beam on this stable polymer film of yours." So it was a fun project, and we learned a lot of physics and chemistry and published lots of papers and filed several patents, but in the application community it didn't go anywhere. Clearly, our particular research setup at that time didn't allow us to answer all the many legitimate technology questions that needed to be answered to be able to put it in perspective with regards to other existing approaches. We knew nothing about costs in a plant environment. Nowadays, the guys in the Research Division are much better informed and equipped to take such ideas to a point where the applied guys have to listen.

HOLLOWAY: We've got something that works already; why should we work with yours?

KAY I think I've talked more than enough technical stuff. What do you think, Paul?

HOLLOWAY: It's your choice. I would like to ask you if you have names of people that you recall who were in the AVS? John and Harold...

KAY: Oh, yeah. I knew lots of people in the AVS, for example, I knew Peter Mark very well before he died and Charlie Duke and Ted Madey and John Yates and Paul Readhead and Gert Ehrlich, Michael Grunze etc. etc., and many, many in the Thin Film and Plasma community like John Thornton, Steve Rossnagel and Jim Harper, Joe Greene and Ivan Petrov, Lars Hultman, Ulf Helmersson, Ludvik Martinu, Jolanta Klemberg Saphieha, Gottlieb Oehrlein, Karen Gleason, Stan Veprek, Richard Gottscho and Vince Donnelly and, of course, Fred Wehner, Dorothy Hoffman and Dick Hoffmann ,and many more.

Several of the people I hired in our lab certainly became part of AVS. People like Dick Brundle, Dan Auerbach, John Coburn , Harold Winters , Alan Luntz, Paul Bagus, Joe Gordon. Who else can I think of? I mean, there was a whole slew of people in our lab. 

HOLLOWAY: You had some theorists that were very active. 

KAY: Oh, yes. By the 80s I had Paul Bagus, who was a quantum theorist and Inder Batra , a band theorist, and Mike Phillpot a surface plasmon expert, and Farid Abraham a molecular dynamics expert, and also John Barker in theory of liquids, and Doug Henderson in double layer elctrochemistry , who all, I think, gave papers occasionally at the AVS over the years. In those early days, one thing that I will say with pride, I was really one of the early people in our lab who said, "We need theoreticians to sort of add a degree of rigor to what we're doing." And they did very well in their respective areas. However, they didn't contribute much to our thin film or plasma projects but certainly to our surface science and solid state physics projects. For example, they did molecular-orbital calculations on how does the of the number of atoms in the electronic structure of a small cluster change as a function of atoms in it , and various other things that were relevant to various sorption phenomena or for example to interpretation of fine structure of core electron spectra, etc.

Obviously, with the availability of the latest computational facilities at IBM it was clear that we could become leaders in many evolving areas of computational science. Very early on, I recruited Enrico Clementi, a quantum chemist, to join us at IBM, San Jose and he then set the pace for this computational effort. 

HOLLOWAY: It goes on and on.

KAY: Oh, yeah. Many good people.

HOLLOWAY: Let me ask - you know, my recollection - and I want to see if your recollection is similar - was that plasma science and technology division of AVS started off in the TOCOMAC area, but that had a limited future.

KAY: In my recollection, absolutely not. It was primarily because of the growing interest and many activities in thin film technologies in various plasma environments..

HOLLOWAY: They didn't have anything to do with thin films, but they looked around and said, "What does this have to do with the thin films and reactive ion etching and plasma deposition stuff?"

KAY: Yes, I think that was how they felt they could make an impact on the existing plasma issues within AVS.

HOLLOWAY: So it became very synergistic at that point.

KAY: I think that became the reason why AVS picked them up. The TOCOMAC people knew something about plasma physics, but - in my recollection, they had zilch to do with starting the AVS Plasma Science Division. I mean, when they lost their jobs, they were happy to learn that energetic particles, albeit in a low-energy regime and plasma diagnostics were of interest to lots of people in the AVS - namely us plasma guys. Their experience was far more in the much, much higher energy regimes, and, of course, had totally different motivation. So I think all the plasma guys and possibly the ion implantation guys were links for them in AVS and their new jobs. By now they clearly are an integral part.

HOLLOWAY: Another area that I was interested in inquiring about was the reviews for the Journal of Vacuum Science and Technology.

KAY: Yeah, well, now that I am no longer doing research --that's what I do.

HOLLOWAY: And you just deal with potential review authors?

KAY: Oh, absolutely. I mean, I'm not doing any research myself any more. I'm not working anymore. You know, I'm 81 years old. So I've finally come to terms with the idea, as I always say, that the world will continue to rotate whether I lubricate it or not. Some of us take longer getting to that point, but I have now arrived. However, I like to keep my foot in the door and follow what new things are evolving in science and technology. At this point, in my age group, the only thing we can offer is sort of a broad perspective. That's really our main, unique advantage that a young person can't have yet. We may not be up-to-date, and probably we're not, and we're not likely to have the most important creative ideas anymore either. But I think we are uniquely suited in evaluating constructively, the worthiness of new ideas and especially, how they fit into other things.

It's now ,what, 15 years since I took on this Critical Review responsibility for JVST. Actually, it was Steve Rossnagel who originally asked would I like to take over this editorship, and it suited my needs very well, because it would certainly allow me continue to keep up with what's going on in the world of science and technology. And another thing you learn when you get older, that it's easier and easier to get a great amount of satisfaction in seeing what other people are doing. When you're young and under lots of pressure, it's usually, "Hey, it's my name on the publication," and I think it takes quite a while to learn that you can have a lot of fun just watching other people succeed, and obviously, at my age, that's really the main thing left in this aspect of our lives. So I started this critical review bit. I shouldn't say I started it, but agreed to take it over.

[Side B]

KAY: I think the way I view my responsibility is to try and keep alert to new, evolving fields of science and technology of interest to the AVS, and then select qualified authors, to share their perspective with the rest of us by writing a critical review of their topic. I found that relatively easy to do.. I mean, all you really have to do is watch the invited paper roster and sort of see what the society as a whole considers as worthy things they want to hear about. Now, obviously, I'm not qualified to decide who are the real experts in all these different technical areas, but the people on the various program committees presumably are. So I go to invited talks primarily, but then, of course, I do make my own assessment, like did it sound exciting? Is it something that other people are likely to be wanting to learn about? But then I have to decide, does a particular speaker communicate well? If they give a lousy talk that puts up a red flag. If they give a very good talk, that's a pretty good sign. 

HOLLOWAY: That's a pretty good sign..

KAY: It's usually not all that hard to get invited speakers to acknowledge that there is probably no one else as well-suited to give us the benefit of their perspective of their field. You would be amazed how many 'egomaniacs' will go along with that. So that's easy, but then comes the hard part. Would you be willing, in fact, to write a Critical Review, but a critical review that doesn't just talk about your own work but puts into perspective other key contributions to the topic and maintains a tutorial note throughout, and also identifies key unresolved science or technology issues. You would be amazed by how many people, who had just agreed that they are indeed the only people qualified to do this topic justice then sort of say, "Well, you know, I'm really pretty damn busy, and also I have to look after my grandmother etc.etc. And then, the few that remain who are eager to do it - and it's usually young people, because they obviously appreciate the exposure and prestige aspects, are usually totally swamped with trying to keep ahead with their research, etc. Then, once they have agreed and made a formal commitment, then comes the worst part, and that's bugging them until they finally come through. Nobody's has ever come through on the originally promised time schedule. It's always anywhere from six month to three years later. And then, of course, comes the same stuff, finding the appropriate referees and subsequently bugging them to come through with their commitments. And you know, these reviews are usually quite lengthy, and so they're a lot of work. In all of this you learn quite a bit about the science involved but also about the authors and referees. And naturally the author after taking maybe several years to come through with it, wants to know where is it? "How come my thing isn't published yet?" Can't you pick more responsible referees?

HOLLOWAY: I gave it to you at least six days ago... [laughter] 

KAY: Okay, that's what I do for JVST. And much of it is a lot of fun. It's a lot of fun, and I kind of hope that it's worthwhile, because I spend a lot of time on it. For somebody like me, it's really my best way to kind of keep up with things I love to hear about. 

HOLLOWAY: Well, you've covered a lot of territory. I was wondering if you had anything else you would like to add at this point.

KAY: Well, I think the AVS has been an incredibly potent science and technology society, and obviously, I'm not the first to say that. I think in contrast, say, to the APS, which, of course, emphasizes physics, but really isn't that committed to worrying about applications. And so you hear some very good science at an APS or an ACS meeting, but the AVS really has the whole spectrum. I think we really do that well; we get some first-class, rigorous science talks, and then we get a lot of competent people who know how to absorb that information and essentially tell us about what the implications are, to technological issues that have to be resolved. 

HOLLOWAY: The AVS, I think, is quite good at cross-pollinating.

KAY: Oh yes. One other thing that comes to mind, which is a huge fairly recent change within the AVS, I mean, I can remember when I knew very few surface physicists who would take the time of day to talk to biologists and vice versa. Today - it's partially because where the funding is, but it's a lot more profound than that. I mean, I think the surface physics world has learned to feel more comfortable with dealing with systems that have more than ten atoms in them, and that together with the incredible progress in surface and interface and nanoscience has allowed them to make a real impact on the bio sciences and in turn, the bioscience world has learned to take advantage of all the sophisticated ways of looking at the world at a more microscopic level. I think the AVS is very clearly making an impact in the bio world, and, I think that's a fantastic thing to watch. It's really exciting, I think.

HOLLOWAY: It's always interesting for somebody of your and my age to see such a dramatic shift. 

KAY: Yeah, compare that with a guy who said, "I'm not going to put CH4 - which actually, like argon, is as dead as a doornail, it doesn't chemisorb on anything at room temperature -  "I'm not putting that in my vacuum system because it's got carbon in it." That same person today (and I'm probably making this up) is probably perfectly willing to look at some macromolecule like DNA, which has more than one carbon atom in it. 

HOLLOWAY: It's a transition that's really interesting. 

KAY: Oh, but it's very exciting, I think. Another thing, and that's my final comment, in this 'nano business'. I really think everybody's aware that there's a lot of hype about it and for very good reason, now that we can study things starting with one atom at a time. But I really think the AVS should find a way of defining what "nano" really applies to, and try to get people not to misuse the term. Possibly, 'any phenomenon or any device that involves isolated units that are less than 100 nanometers', qualifies. It should either uniquely depend on 'surface to volume' effects and/or 'quantum effects'. Everything else is really just a natural progression of beautiful work that many people have done for many years, like the work in catalysis, for example. They get sort of offended to be told, "Hey, we've got a whole new thing," and they say, "Well, what's conceptually so new, except that you can make things smaller and look at them and have found new applications.

I think eventually it will get there. I've heard a lot of people get offended by the excitement that somebody arouses when he calls it 'nano', and then by implication, the stuff that you guys did years ago, that's all old-fashioned crud, and conceptually, it really isn't.

HOLLOWAY: That's true. Okay.

KAY: Okay, thank you for this opportunity. And as usual for me, it was brief. [Laughter]

return to top