AVS Historical Persons
| Paul Redhead - 1991
Paul Redhead - 1991
Oral History Interview with Paul Redhead
Medard W. Welch Award Winner, 1975
Interviewed by Bruce Kendall, Feb. 12, 1991
: I'm Bruce Kendall. This interview is part of a series being made for the American Vacuum Society historical archives. It's February 12, 1991. We're in Clearwater, Florida. I'll be talking today with Paul Redhead, Researcher Emeritus at the National Research Council in Canada, former president of the American Vacuum Society, editor of the Journal of Vacuum Science & Technology at one time, former board member, author of the first major book on ultra-high vacuum. I could go on and on, but we'll let Paul tell us about it in his own words. What were your reasons for getting into ultra-high vacuum in the first place, Paul?
: Well, really as a means to an end, as far as I can recollect now. This was in the mid- to late-'50s. We were planning to do work on surface work, what would these days be called surface science. In particular, I was interested in examining some elementary chemisorption reactions on metals. We knew of the pioneering work of Alpert and his group at Westinghouse that produced ultra-high vacuum pressures down in the 10-10 Torr range. These low pressures, of course, were necessary to maintain surfaces in a clean and reproducible state for a reasonable length of time so you could make measurements on them before the characteristics had changed.
At that time, there were only two or three groups in the world that had these techniques available. There was Alpert's group at Westinghouse labs in Pittsburgh. There were the people at Bell Labs, Murray Hill, Joe Becker, and several people at Murray Hill. That was about it. I can't recall any labs in Europe that had the technology, though maybe my memory is at fault.
We set about trying to establish the techniques in our own lab. Since we had very little money, we had to make everything ourselves. In any event at that time, the only ultra-high vacuum apparatus we could buy were Granville- Phillips valves, which had just come on the market. Everything else we made ourselves, glass systems.
The first thing that we attempted to develop were cold cathode gauges. The notion was at that time, cold cathode gauges had been in existence since I suppose about 1935-36 when Penning first demonstrated you could maintain a discharge at low pressures and crossed electric and magnetic fields. There was a mythology around in the fifties that cold cathode gauges always went out. The discharge actually went out at pressures below, say, 10-7 Torr or something of that order. I could see no reason why this should be so, no theoretical reason or conceptual reason.
So we built some gauges in which we deliberately tried to trap the electron cloud as efficiently as possible and tested them. Lo and behold, they didn't go out. This work was largely based on considering the work that Haefer had done in Austria. Haefer had just published in a journal called Acta Physica Austriaca. He�d done some very nice work in studying what we later called inverted magnetron structures. But he hadn't gone down to ultra-high vacuum conditions.
We attempted to see whether we could extend Haefer's work, and very quickly we found that yes, we could get these gauges to work down certainly into the 10-9 to 10-10 Torr range. The first were inverted magnetron gauges and then quite accidentally one day, I'd made an inverted magnetron gauge in which the end plates had separate electrical connections so I could measure currents through them. It just occurred to me to connect up the other way around, which I did. It created a magnetron gauge. Somewhat to my surprise, the sensitivity or the current-per-unit pressure went up by a factor of five.
That was the accidental way in which the magnetron gauge came about. In due course some years later, a tidied-up version of the magnetron gauge went to the moon. I can't remember which shot it was 17 I think? I'm sorry, I can't remember. But anyway, it went to the moon to measure atmospheric pressure on the surface of the moon.
: You can tell us a little bit about what pressures were measured on the moon.
: I wasn't directly involved myself. The actual gauges that went had been developed by National Research Corporation in Boston. I think the interesting results of those measurements on the moon were the sudden increase in pressure on the sunrise on the moon's surface due to out-gassing, presumably, of the moon's surface. That was a bit unexpected. I can't remember how high the pressure went at these sunrise bursts, but it was quite high. I think it was in the 10-6 Torr range. It rapidly came back down.
The development of the cold cathode gauges led us to a whole set of techniques to do "cheaparino" ultra-high vacuum measurements. You may remember yourself at the time when you joined us as a post-doctorate fellow that we had the lab full of these little cheaparino ultra-high vacuum systems which, again because we had very little money, we had to make these things cheaply. My recollection is that they cost us a good deal less than $1000 per system. But basically the only part of the system we had to buy was the Granville-Phillips valve. They were pumped by little pumps that we developed from the magnetron gauge, which had titanium evaporators in them. There were glass systems, more glass systems. We used to pump them down from atmosphere with adsorption pumps or vacuum pumps and an external ion pump.
So we had little units that you may recall we called 'Fidos' because they were a little unit on casters that could be run around. This, again, to save money so you only needed one of these units to service several of the ultra-high vacuum units. Because once we had baked the systems with the external pumps on, we then closed the main ultra-high valve and pumped with these little internal magnetron pumps and didn't need the external pumps any longer. We had, I think, two or three Fidos, as we called them, to service well over a dozen ultra-high vacuum systems.
The systems were sufficiently cheap and easy to make so that one just threw one together for each experiment that you wanted, rather than having-- It was a cheap approach that was convenient at that time rather than the modern approach of having these vast stainless steel bagpipes with teats all over the place and moving gadgets, which are very nice, but also very expensive.
: The bake-out temperatures were quite high by modern standards, weren't they?
: Yes. We used to bake out at 450C. There were occasions when, number one, when the thermocouple fell out of the oven, so the controller didn't switch it off and the temperature went up. When you came in, in the morning, the whole of the ultra-high vacuum system was about one-centimeter thick melted down on top of the...
: Then later you got onto some electron-stimulated desorption work.
: Yes. That, again, had an accidental start. Nearly everything I've done, I sort of fumbled my way into by accident. I was trying to measure the change in work function of the tungsten surface as it adsorbed oxygen. This was a contact-potential measurement, various gases. When I tried it with oxygen, I was very puzzled by the response of the Bayard-Alpert gauge. You would put in this oxygen, and the pressure reading of the Bayard-Alpert gauge would go up to sometimes as much as 100-times higher than the indication of the cold cathode gauge that we had on the system. It would stay up there after you'd pump the oxygen away. It was very puzzling. I fiddled around with this for a month or so trying to understand it.
My first thought was that what was happening was that the X-ray production was being modified in some way by the presence of absorbed oxygen. In fact, I published a little note in Vacuum to that effect, which was entirely wrong. I now try and pretend that I never published that note. But after some month or so of experiment, it became obvious that, well, not obvious, but it became fairly clear that what was happening was that the oxygen was adsorbing on the grid of the Bayard-Alpert gauge. The electrons were desorbing O+ ions. It was this that we were measuring.
Why we were able to disentangle this problem was that the Bayard-Alpert gauges we were using were modulated gauges, which we had developed by this time. It was, again quite accidentally, for reasons that would take too long to explain now, when you switch the modulator potential on a Bayard-Alpert gauge, it has very little effect on these ions that come from these electron-desorbed ions. The result is that one can measure, by switching the modulator, this electron desorption quite easily. It was that that let us disentangle it.
Then I got interested in the process, and studied the process in some detail. Again, an example of the sort of happy accidents that occur in science, just as I was getting prepared to publish this paper on electronic desorption, as we called it then, of oxygen, I was down in Chicago visiting Bob Gomer and discovered that he had just found exactly the same thing by an entirely different method, using field emission, field electron microscopy. He was just preparing to publish. We exchanged manuscripts, and we both published, I think, in different journals within a month of one another. The effect, or at least the explanation of the effect, is known as the Menzel-Gomer-Redhead Theory. It was, again, an example of something that happened so frequently in science, these coincidences which aren't really coincidences; they result because the techniques and the sort of mindset is all available. Things get invented in the same place and different places at the same time.
: You had a whole collection of elegant gauges, a modulated Bayard-Alpert gauge, suppressor gauges, I think extractor gauges. Can you tell us a little about the work on them?
: What happened was, these ultra-high vacuum systems that we used in the group could achieve pressures routinely of a few times 10-11, perhaps three or four times 10-11 Torr. It was that it was a little messy to measure, because although we could measure it with the modulated Bayard-Alpert gauges, it was kind of unsatisfactory not being able to do a direct measurement of it. The cold cathode gauges could measure it, but their stability and linearity wasn't very good.
So we started trying to develop gauges with lower X-ray limits. The first one that we did was a copy of the suppressor gauge, the principle of which had actually been first demonstrated by Metson back in 1951. Metson at the British Post Office Research Station in England invented the suppressor gauge the same year that Alpert developed the Bayard-Alpert gauge. But the Bayard-Alpert gauge was so much more simple and elegant that that took over, and Metson's gauge was never followed up. Then some years later, de Segovia, who is here at this meeting and was working with Alpert at Westinghouse as some sort of a visiting student, I believe, developed the suppressor gauge. Well, we copied it with modifications.
We found problems (again, I don't have time to go into it), but it turns out the soft X-rays reflect with a high probability you aren't perhaps aware of this, that the soft X-rays in the sort of 100 Volt range tend to bang around due to reflection with reflection coefficients as high as 30%. So they do go around corners. Suppressor gauges are not quite as good as you expect them to be because the X-rays go around corners.
Finally, after a series of gauges, we developed the extractor gauge, which did push the X-ray limit down into the 10-12 Torr range. There's a commercial version of that gauge available from Leybold and has been for a long time.
: You led a large research group for many decades, had a lot of interactions with other research groups in many countries. Would you care to comment on some of the personalities you've come into contact with?
: I've been very fortunate in my colleagues and very fortunate in working in a laboratory in which, at least in the period we're talking about no longer true, unfortunately we had the most extraordinary freedom to choose the research we wanted to do. Admittedly, we had no money to do it with, but that, in many ways, was an incentive to be smart rather than solve your problems by building more expensive equipment.
It was a lot of fun. Over the years there were several people who joined us as post-doctorate fellows some of you gentlemen from the Antipodes, such as yourself and John Robins. We had a couple of Japanese gentlemen, Ishikawa, I don't know whether you overlapped with him, and so on. So we had a steady flow of people, many of whom have ended up in areas related to the vacuum technology business since then. I was very lucky in my colleagues of Pete Hobson and Ernie Kornelsen. The three of us put together this book on ultra-high vacuum.
We attempted to divide up the areas of work directed towards improving our general technologies for ultra-high vacuum, and also trying to understand the basic mechanisms involved. Ernie Kornelsen, for instance, spent many years studying in great detail the way in which ions are trapped in solids, which started out as a piece of purely technological work to develop ion pumps and understand them. But it turned into some very nice fundamental work in understanding the entrapment of rare gas ions in solids and the discovery of channeling and so forth resulted from that.
May I tell you a little story? I'm reminded of how the modulator gauge came about, which again was entirely by accident. This was when we were just beginning to make our own ion pumps and trying to understand how they worked. I had been doing what became known as thermal desorption spectroscopy with chemisorbed species. I suggested to Ernie Kornelsen that maybe we could do the same thing, but instead of chemisorbing the gas on the surface, we could shoot ions into a metal and then heat it up and see what can pop out again. So we did a Friday afternoon experiment to see whether this would work. The quick way of doing it was to put a hairpin filament inside the grid of a Bayard-Alpert gauge, which we did do very quickly. We then operated the gauge and drove the ions into the filament for a good length of time, then pumped the gas away, connected things up so that it was acting as a gauge again and heated this filament up to watch the pressure rise caused by the gas coming out of the filament.
The experiment didn't work very well for obvious reasons, but it showed that the concept was feasible. The work that came out of it was, in this process, we discovered of course that by switching the potential of this extra filament, we could modulate the ion current to the main collector. The penny dropped at this point. Then we just made a modulated gauge. We used them ever afterwards in very convenient gadgets. But I suspect that more advances in science happen accidentally like that than most of us care to admit.
: I felt that one of the main unique things about your lab was that it showed a proper awareness of lab safety issues. This was a time when lab safety issues were not fashionable. I wonder if you'd like to comment on the famous mercury event.
: [Laughs] Yes. And you were involved up to your neck in that one. Very briefly, what happened was that we were still using mercury diffusion pumps at the time. We had copied a design of a glass diffusion pump that they used at Bell Labs. It was a beautiful pump, you know. A glass blower had made them. It was a very good pump. We had the cooling water for this pump, of course, running in rubber hoses. I don't quite remember how the hose broke. I remember we had a bird come into the lab. It came in through the opening from the fume hood. And the door of the fume hood was open on the inside, and it flew into the lab. I'm not sure whether this was the same event. Anyway, the bird got itself between the belt and the pulley on the vacuum pump and threw the belt off the pulley and in the process broke one of the rubber pipes. The jet of cold water hit the boiler of the pump, cracked the glass. The boiler fell down onto the heater and of course, all the safety devices had been jimmied. The heater stayed on. The boiler itself didn't crack. So the whole boiler full of mercury was evaporated over the lab.
I came in on a Sunday to do something. Fortunately there was a glass panel in the door of the lab. I looked through this glass panel, and the whole lab looked gray. Fortunately I didn't open the door. I stood there and wondered, 'What has happened?' Then I realized it was probably mercury and went away and got a gas mask and went in. Sure enough, the whole lab was covered with mercury. I remember with great satisfaction the lab was air-conditioned, so the windows wouldn't open. I remember smashing the windows with a fire axe which was great fun and phoning everybody. Everybody came in on a Sunday. God help us, the thing that saved our lives was you, because you had been involved in some sort of mercury spill in Australia. When you arrived, you insisted that we buy a mercury vapor meter. So I went out and ordered the mercury vapor meter, so we had this thing. Otherwise we would never have realized the level of mercury vapor.
As you recall, I think it took us the best part of a week to sort of clean things up. It was months before we were back working again. The final irony of this situation was that the lab was right over the kitchens for the cafeteria. They were on the same air-conditioning system. I was terrified that all the kitchens were contaminated, but fortunately they weren't.
: Exciting days. I noticed that a couple of your papers are still being cited many decades after they were published. I think one of them is the thermal desorption paper. Tell us a bit about that.
: Yes. I don't know why it's still being cited. I suppose it's to some extent a how-to-do-it paper, which citations do tend to go on. But it was Ehrlich who was at the time at General Electric Schenectady who did really the early work on thermal desorption spectrometry and worked with a guy by the name of Tom Hickmott. He did some lovely work. The only thing that I did was to sort of extend Ehrlich's work and increase the pumping speed and lengthen the time of the thermal scan. Ehrlich was doing the thermal scan in seconds. The result was that his pressure curves didn't decrease, if you follow me. I'm not explaining this very well. What I did was to lengthen the thermal scan to about a minute with a fairly high pumping speed. The result of that was that you got the pressure was the derivative, if you like, of the rate of accumulation of molecules in the gas phase. You got peaks associated with each activation energy of desorption. That was a contribution I made. But that's a technology which has gone on and is still very widely used in the study of surface phenomena. The other paper that still gets cited is the paper on electronic stimulated desorption of oxygen.
: What is it that you're working on now?
: I'm trying to do what my mother told me she tried to do in the latter part of her life, which is to grow old gracefully. I'm not being very successful at it. Since I retired and have escaped the administrative and managerial duties, I've done some work looking, once again, to try and understand cross-field cold cathode discharges, which are still quite mysterious in understanding their instabilities. In particular, I would like to explain the non-linear nature of the current versus pressure characteristics, which no theory can explain at the moment. I think I know what causes it, but I can't explain it mathematically. So I've been interested in trying to sort out some of those problems.
: Thank you, Paul, for taking the time to be with us today and for making this contribution to the AVS historical archives.
: Thank you very much, Bruce. It's been a pleasure.