Interview with Ivan Stensgaard
CAMP Aarhus: The Scanning Tunneling Microscopy Group, Aarhus University, Denmark, 8 March 2001
Arne Hessenbruch: But during the period of '82-'87 you would have known about the STM?
Ivan Stensgaard: Yes.
Arne Hessenbruch: And it was something that you perhaps watched? You watched the development of this instrumentation and you were critical to see whether it would prove its mettle for you to begin to use it?
Ivan Stensgaard: Yes. Well we were not, as I recall it, completely convinced that it would add a lot to the insight into metal surfaces. But as soon as we did enter the field ourselves, placed the instrument in vacuum and seen the results, we were immediately convinced. I'd say from then on it has been fantastic: it has been quite easy to achieve good results with just about any topic of investigation. One saw new phenomena almost regardless of what one studied--new and completely unanticipated phenomena.
Arne Hessenbruch: So you're saying "we" now, this is presumably the CAMP group. When did that gel? Was that in '87?
Ivan Stensgaard: Yes, I do mean the CAMP group. But we began collaborating earlier. Historically, I think it was like this: I returned from the US and received this grant for building up an ion scattering facility for studying surfaces [referring to an earlier mention that is not on the tape this part of the interview will be repeated in the near future]. Some years later, Flemming [Besenbacher] and I worked together on ion scattering from surfaces. We looked at small molecules, and specialized on the examination of hydrogen on metal surfaces. Then in the 80's, probably around '85 or so, we began to discuss in more detail this question of building a scanning tunneling microscope. Flemming took the initiative in those days and contacted Erik Lægsgaard. Erik was really the key person in building the STM here. And from when he started until he had a table-top version, giving atomic resolution on graphite, lasted only a few months. Of course it took longer to make an instrument fitting into a vacuum system, and so on.
Arne Hessenbruch: Did you make it locally?
Ivan Stensgaard: Yes, everything here has always been homemade
Arne Hessenbruch: You've never bought instruments from the outside?
Ivan Stensgaard: No. We tend to claim here that our instrument is better than what we can buy. It may sound like boasting, but I think our results help prove the contention. The fact that we can work with extremely rudimentary vibration isolation, proves that we have an highly stable instrument.
Arne Hessenbruch: So what were the characteristics of the design the Aarhus design?
Ivan Stensgaard: Well, there are two basic ingredients in the construction: one is the scanner tube scanning the tip across the surfaces, and I think that was copied simply from the Binnig and Rohrer design.
Arne Hessenbruch: The empty cylinder.
Ivan Stensgaard: Yes. So, the larger problem was how to approach this scanner tube tip to the surface in a controlled way so that you would come close enough to draw a tunnel current without actually colliding. To begin with we had many ideas about some mechanical way of doing it with micrometer screws, but soon we focused upon piezoelectric materials. And I think, again, that it was Erik's idea to make an inchworm. This is a really strange story. We had a very small inchworm made: only a few millimeters in diameter - maybe 10-12 mm long. It worked well and a company called DME in Copenhagen wanted to commercialize our setup. During the discussions and negotiations with them, they found out that the inchworm principle was already known, and what was worse, it was already patented. That was a something of a setback - not for our in-house microscopes but for the efforts to commercialize it. The commercially available inchworms were much bigger and not nearly as elegant as ours. But the principle was patented, and the patent had a few years left before expiration.
Arne Hessenbruch: Oh, so the patented inchworm was not intended for the STM?
Ivan Stensgaard: No. The patent was about 10-15 years old and had been developed for something else. Whether it was developed for one of the topografiners and what they were called, the instruments used prior to STMs, I don't know. We can check the patent.
Arne Hessenbruch: Well, the topografiner was invented at the National Bureau Standards [now NIST]and I don't think they would have taken a patent.
Ivan Stensgaard: We could check it.
Arne Hessenbruch: So you designed the inchworm in '87, just when you were making your first STM?
Ivan Stensgaard: Yes.
Arne Hessenbruch: But your first STM did not have an inchworm.
Ivan Stensgaard: Yes, it was a mechanical approach.
Arne Hessenbruch: Where did you get all the information for making an STM? Did you look at the publications of Binnig and Rohrer, did you go and see them? Did you go to the IBM labs?
Ivan Stensgaard: No, the way I recall it, we had no personal contact. I think everything was done just using the publications.
Arne Hessenbruch: And it wasn't particularly difficult?
Ivan Stensgaard: Not if you had a clever man like Erik. The short time required to make the first STM also indicates that it was fairly easy. Of course in retrospect many aspects of building an STM seem easy, because we know just how piezos work and what to pay attention to. The game nowadays is really to make them as compact as possible, to make them work even with vibrations in the surroundings and so on.
Arne Hessenbruch: Yes, you were saying you had very elementary vibration damping. This is another characteristic of this design?
Ivan Stensgaard: Yes. Exactly. When you visit other STM labs, you're frequently told that they have to work at night when there is no traffic. They have all kinds of vibration damping, entire vacuum systems hanging in rubber -- springs from the ceiling or something like that. They have the microscope damped with magnetic or eddy current damping. And people had to stay quiet and not move about in the room and so on. That remained characteristic until at least ten years ago. By contrast, from the beginning we hung our microscope on three or four springs, or viton-o rings, which are really rubber springs. That's all; we have no further vibration damping. And people can even touch the UHV chamber without interfering with the tunneling. So that has been the main characteristics of the instrument here: its ability to scan fast and under adverse conditions.
Arne Hessenbruch: Fast and robust. You have told us about the robustness, could you tell us about the speed?
Ivan Stensgaard: Fast scanning is interesting also because it has opened up new opportunities compared to other microscopes. There are more fast-scanning microscopes today but to begin with we scanned faster than most others. We scanned at the rate of one second per image, where many others spent one minute per image. This gives you the opportunity to follow dynamic processes on surfaces by sequentially recording images and then replaying the images at a faster rate - in the form of a movie. That gives you a very vivid impression of what happens on the surface. You might object that a movie yields no new information, but it turns out to be a very good means of attracting attention to dynamic processes. Such phenomena miss easily when just examining static pictures. The eye is well-suited to spotting change, so that even when there is a lot to keep track of, say 20 to 40 entities on a surface, your eye will immediately pick it up a movement of any one entity. So this question of scanning fast and taking movies, is not just a question of presenting results in a nice way, it also gives you a better insight into what is going on.
Arne Hessenbruch: But you weren't capable of making movies off the bat in 1987 were you? How long did you need to scan in 1987, roughly?
Ivan Stensgaard: I don't recall it really. I think we could record images in seconds, but the problem was probably rather that saving and replaying images requires a lot of data storage, and in those days the hard disks were not very big. It was slow process simply to get an image from the hard disk onto the screen.
Arne Hessenbruch: Was your ability to scan quickly a desideratum in 1987?
Ivan Stensgaard: Early on we probably didn't focus on the time you spent recording. It probably didn't matter whether it was 1 or 20 seconds or so. But I see now [checking on the computer] that taking movies and things like that, must have started in the late 80's. Because we have publications at least from '91 where we have included, not movies because that wasn't possible, but results based on these techniques.
Arne Hessenbruch: And what sort of thing did you image? Metal surfaces with adsorbed atoms?
Ivan Stensgaard: Yes, of course the standard sample in those days (as it still is for testing an STM in air) was graphite. Of course we used that as the test that our STM was working properly. Having done that we moved rather swiftly in the surface science direction again. We examined some gold-plated plastics, simply to investigate the micro-roughness. The purpose was to test the coating of large mirrors of an x-ray telescope to be used in a satellite. We also investigated some biological samples. But we were really implementing the STM in a vacuum.
Arne Hessenbruch: You say you were immediately convinced, when you saw the images, that this was a useful tool. It was seeing the images that convinced you almost instantly?
Ivan Stensgaard: Yes
Arne Hessenbruch: Why? There are many problems with the imaging process of course, with the noise and the interpreting just what the tunneling current actually is referring to. On metals this is less of a problem than semiconductors. Was that a help to you?
Ivan Stensgaard: Yes, to some extent it was. Let me just mention the first system we examined, to maybe make references to that. Once we got the microscope into a vacuum we continued our work on small molecules on a metal surface. The first system we examined was one I had investigated earlier using ion scattering. We had absorbed oxygen on a particular copper surface, a copper (110) surface. And it was known what happened to the oxygen, how the surface reconstructed. What was not known was how the process took place, and that was revealed immediately almost with the STM, because we could look upon the surface while the reconstruction took place, and we could make a movie and watch the process.
Arne Hessenbruch: That must have been exciting.
Ivan Stensgaard: Yes, and then I was at least totally convinced at that time that here was an amazing tool.
Arne Hessenbruch: So the problems that there might be with interpretation were really secondary to you, you said that even with all those problems, "it's still exactly what I need?"
Ivan Stensgaard: Yes, you said that in the case of semiconductors it is more difficult to relate the STM-produced image to the surface corrugation. That is correct of course, but you could say that we had a similar problem on our hands with oxygen on a copper surface: were we imaging the copper or the oxygen? And it turned out that we were usually imaging the copper. This posed the question: why not the oxygen? And why is it sometimes just the other way around, that we image the oxygen and not the copper. So scanning on metals was not completely free of problems, at least not when you have the adsorbates on them. But if you image one clean surface with only one element present, then the interpretation is straightforward.
Arne Hessenbruch: What I'm trying to understand is that when you use a completely new tool, you cannot know whether it works.
Ivan Stensgaard: I think the problems of instrumentation were so insignificant compared to the insight gained, that we were absolutely convinced of the amazing utility of the instrument. That's the way I recall it, really. There was a trend in those days--probably through the 90's--from static to dynamic problems. Other methods, such as low-energy electron diffraction, high-energy ion scattering, x-ray diffraction, are better suited for static problems. You can induce a reaction on the surface, and then you can take your tool and find out how the surface has changed. But with the STM you had a tool with which to see the transformation from one state to the next. Later on, you could even follow single atoms moving on the surface. That was really amazing, and almost impossible with other techniques. Field ion microscopy has been able to do it in some cases.
Arne Hessenbruch: Before the STM, you were using tools like ion-scattering, but it was not the tool that defined your research. Now it looks like the STM defines the group.
Ivan Stensgaard: I don't really think that's correct. No matter what you investigate, you have to take both things into account. You have to find an interesting problem amenable to investigation with your background, its relevance and so on. Then you have to consider the available techniques. You have to find a topic amenable to the techniques available to you. I think that also describes what we did with the STM. It is a question of resources. The microscope does not define the activities; a natural development has rather taken place. I mentioned before that we first investigated small molecules on surfaces, and we have continued to do that. Our research has now diversified but it is still to a large extent coupled to chemical reactions and the like. We have studied surfaces in greater detail also clean surfaces although clean surfaces are less interesting for STM research because one can get the information by other means. But when you grow metal on metal for example, what happens? How do these atoms of one metal, influence the properties of a surface of another metal. How does the surface restructure? And, to come back to my initial point, how do the chemical properties change? What happens if we now adsorb oxygen, what happens if we carry out a chemical reaction, let's say oxygen and CO reacting on the surface. In all these cases there has been a connection to the initial questions.
Arne Hessenbruch: Have you expanded the number of chemical reactions you worked with?
Ivan Stensgaard: Over the years we have worked with a number of chemical reactions, but in the last few years we have focused more on catalysis. We have for many years collaborated with Haldor Topsøe AS, one of the leading companies in the catalytic world. We have collaborated on a number of issues, in once case a new catalyst was developed on the basis of our results. We proved scientifically that the catalytic conversion works. Whether that catalyst in the end will be commercially available now depends on other considerations, such as the cost of producing a catalyst, its lifetime and so on.
Arne Hessenbruch: But you discovered new catalysts with the STM, is that what you're saying?
Ivan Stensgaard: We discovered a system that turned out to have possibilities in catalysis. We could talk in more detail about that, but it's a story in itself.
Arne Hessenbruch: Okay. What can you do now with the STM that you couldn't do in 1990?
Ivan Stensgaard: There has been constant improvement in speed: the rate at which you record and replay images. And new designs have made the microscope usable under new conditions. It now works over a large range of temperatures and pressures. To begin with experiments could only be carried out at room temperature, whereas we now can work from 20K or so and up to maybe 150ºC (500K). This increase in the temperature range is important for the investigation of dynamic processes because the rate at which they proceed is governed by temperature. And if you wish to study a certain reaction or a certain diffusion property, you have to have the events happen at a rate you can follow. That means that you shouldn't have too many events per second, preferably no more than one event every few seconds. And you can normally do that by adjusting the temperature.
We can also work at many different pressures now. Early on we had to work either at atmospheric (ambient) pressure or else in a very high vacuum. We have recently become able to work in the full range of pressures from ultra high vacuum to atmospheric pressures and could in principle make the system work at even higher pressures, if it weren't for some windows unable to withstand atmospheric pressure. And that again is advantageous for catalytic research, since all catalytic processes run at high pressure and high temperature. In the surface science community there has always been arguments that their field would help improve catalysts and understand catalytic reactions. That is to some extent true, but there has also been a lingering skepticism whether the processes taking place at atmospheric pressure remain the processes governing chemical reactions while raising the pressure from 10-10 Torr to atmospheric pressure. So that has been an important result: that at least in the one case that we have done so far, the system behaves identically at high pressure and in an ultrahigh vacuum, it's exactly the same things that govern the processes.
Arne Hessenbruch: What do you have to change in the STM to make it amenable to the range of pressure and temperature?
Ivan Stensgaard: Actually not much was changed; it's basically the same setup. The only major change is that it is located in a minor vessel; it's in a minor chamber with a smaller volume, which is the volume that is brought up to high pressure. The suspension system has changed also, but the basic STM is the same. The suspension system is still based on a few springs, and looks differently, but the system as such is really the same. And of course minor things like gold coating surfaces to avoid reactions of those, but the heart of the microscope is completely unchanged.
Arne Hessenbruch: Why didn't you do these things 10 years ago?
Ivan Stensgaard: I think in the field of ultrahigh vacuum there have been so many extremely interesting topics that there simply hasn't been the capacity to look in other directions. We should bear in mind that when CAMP was established in '93, much more funding became available than before. And the group grew much bigger. And although to begin with we also focused very much on processes at low to high vacuum, then at least during the second period of CAMP (from 1998 onwards) we have been able to diversify. High pressure is simply a natural development out of our original research. When I refer to new fields, I mean investigating organic molecules and other biology-related issues. So our research program has really developed quite naturally.
Arne Hessenbruch: So you started CAMP in 1993 - you, Besenbacher and Lægsgaard.
Ivan Stensgaard: The history of CAMP is more complicated it is related to changes in the funding of Danish research. A large insurance company, Statens Livsforsikringsanstalt, was sold and a fund, Grundforskningsfonden, was set up with the money from there: 1 or 2 billion Danish kroners. Interest from the capital there was to be spent on basic research in all fields. In those days the interest rate was high and leading to a fair amount of interest each year. They requested and received many proposals for research centers. I don't recall exactly but it was a very large number. And in the end they funded 23 centers in all kinds of science: medicine, humanities Søren Kierkegaard. We got a center too, a collaboration between our group here at Aarhus and Jens Nørskov's group at the Danish Technical University at Lyngby. The group there focused mainly on theory, and Nørskov was the director of the center, as he still is technically. The reason the two groups joined in this project was that we had had a prior collaboration within another program called FTU, research for technological development. In addition Flemming had collaborated with Jens Nørskov on another topic. So we knew each other. A proposal was made based on the joint resources of these two groups and after some external referring procedures it was eventually funded.
Arne Hessenbruch: And the project was focused on the STM?
Ivan Stensgaard: I think the scope was broader: the interaction between small molecules and surfaces, to some extent chemical reactions. Our tool of investigation was STM but the center as such was not a STM center, it was a center doing research in a specific field. And our tool of research was the STM.
Arne Hessenbruch: You mentioned that you had contacts with Topsøe and other companies. Have you had contact with chemists, people in chemistry departments?
Ivan Stensgaard: We have, but not much.
Arne Hessenbruch: Are they not interested in this?
Ivan Stensgaard: I think they are in many cases, but we work in surface physics, as it is called in Danish. The English term, "surface science", is more neutral. Of course what we do is to some extent surface chemistry. But our background is just physics. You could say that a center like this might just as well have been placed at an institute of chemistry. It just so happens that we had this collaboration, and that we had not had the opportunity to interact much with chemists.
Arne Hessenbruch: So it's not the barriers between physics and chemistry that are particularly high, it just hasn't happened?
Ivan Stensgaard: It just hasn't happened. It might have worked out differently, had we had a common institute instead of a separate physics and chemistry institute. Some of the activities going on here might as well have been placed in the institute of chemistry and vice versa. Nowadays it's not that natural to have this division into two institutes.
Arne Hessenbruch: How far is it from here to chemistry?
Ivan Stensgaard: It's just across the parking lot, 100 meters. We're even connected underground to them, although it takes a specific key.
Arne Hessenbruch: But you don't go to each other's seminars?
Ivan Stensgaard: Very infrequently.
Arne Hessenbruch: This does seem surprising and in need of an explanation. No?
Ivan Stensgaard: I think much of what we're doing is of interest to chemistry as such, but our institute is comparatively small. Chemistry and physics are amazingly broad fields and our small institute is only able to cover certain areas. It has just so happened that we work in a field without a natural counterpart in the institute of chemistry. There are no ill feelings on the other side of the parking lot; it's not based on that.
Arne Hessenbruch: And what about chemists in general out there in the world? Do you get resonances? Where do you publish, for instance?
Ivan Stensgaard: We publish in papers with as high an impact as possible. We have published very large number of Physical Review Letters during the period of CAMP, on very few occasions in Nature and Science also. But that is not that easy, and sometimes you feel it is more difficult being European than if you had been American, but that's another story. We publish of course also in other journals; some material will go to Surface Science. Surface Science does not have that high an impact, but it is obviously a very natural journal for people working in surface science. In my opinion the quality is rather mixed, but these are the main journals that we choose. We do not publish much in European journals, or journals that are particularly European such as European Journal of Physics and European Physics Letters.
Arne Hessenbruch: For any particular reason?
Ivan Stensgaard: Simply because impact and visibility are low.
Arne Hessenbruch: When you say that in Physical Review Letters the visibility is high, is it not limited to the physics community? Isn't your work broader?
Ivan Stensgaard: It's probably correct, but we have colleagues in other places working in chemistry departments. And some of them will also publish in Physical Review Letters. I think it is recognized as some natural place for good work, also. But we have also published in Chemical Physics Letters, there are also a number of publications there, but they're not quite as prestigious as Physical Review Letters.
Arne Hessenbruch: Are you under pressure to produce results of commercial interest and if so, does this conflict with your own interest in pure science?
Ivan Stensgaard: I don't' think so. I don't think we are under much pressure in CAMP to produce results of direct relevance to industry. I do think that we in CAMP all see an extra benefit or satisfaction in producing something of relevance to the "real world and not just to pure science.
Arne Hessenbruch: So when you get around to renewing CAMP for another five years, the grant-giving bodies will not look to commercial applications?
Ivan Stensgaard: It won't come to that. The funding will stop after ten years. It's the general policy that centers such as ours may reapply only once, so that they run for a maximum of two periods. After two periods, they ought to find alternative funding and in some way become a more integral part of the institute. What will happen in two to two and a half years from now, I don't know.
Arne Hessenbruch: You are not worried?
Ivan Stensgaard: Not terribly, I think some kind of funding will come. I think that the government has set aside some funding for integrating centers into their host institutions.
Arne Hessenbruch: The reason I'm asking these things, is that Jens E. T. Andersen at the DTU was telling me about tremendous pressures. The DTU has now a patent office, and renewal of academic contracts is subject to results in terms of patents rather than scientific publications.
Ivan Stensgaard: The emphasis on patents has definitely increased here also, but in my opinion it has not yet become a pressure. There is a greater emphasis on it, and some of that emphasis originates from the government. A new law was introduced in Denmark. Up until maybe a year ago, you had the rights to exploit your own inventions, even as an employee and civil servant. That has now changed so that when you develop anything patentable or otherwise exploitable, you have to first contact the university. Your invention will be assessed. The university decides whether it wants to patent your invention and exploit it. If it declines, then you are free to do it yourself. So you could say that maybe the timing is a little bit strange, because on the one hand you remove some carrots from the researcher and on the other hand you apply the stick to encourage work resulting in patents. I think the general political perspective is that a fair amount of knowledge of interest to industries is generated at the universities, but no good forum for the transferal of that knowledge exists. Maybe the thinking was that having the knowledge documented in patents would facilitate the transferal.
Arne Hessenbruch: Is it possible to pinpoint advances made in the field of catalysis over the last ten years related to the possibilities of the STM?
Ivan Stensgaard: I think this story about developing a catalyst from scratch based on pure research is rare, almost unique. There are no more than one or two other examples. Normally catalysis has been a highly empirical field. Developments have mostly been trial and error - of course guided by different kinds of knowledge. But in the end it came down to trying something, modifying it, changing the stoichiometry or the ratios, optimizing and so on. The understanding of the catalyst emerged gradually out of this process. By contrast, ours is a clear case of going directly from scientific insight into the working of a catalyst to its development.
This page was last updated on 19 July 2001 by Arne Hessenbruch.