Materials Research Activities

Pierre- Gilles De Gennes interview by Hervé Arribart and Bernadette Bensaude-Vincent

Interview of Pierre-Gilles De Gennes by Hervé Arribart and Bernadette Bensaude-Vincent

May 2, 2002, Paris, Ecole Supérieure de Physique et de Chimie industrielles

Hervé Arribart (HA): Some articles published in the USA in 1991 presented you as a materials scientist. Do you consider yourself as such?

Pierre-Gilles de Gennes (PGDG): It depends on the period you are talking about. When we were in superconductors we did not consider ourselves as materials scientists. In fact we had a happy period when we could perform any amusing experiments but when things became more complex we left. For instance we had understood what vortices were doing in classical superconductors. Then there was a second stage where you should invent alloys which had special precipitates so that they would pin the vortices. That sort of action was beyond our technical means (we had very limited means in Orsay). So precisely at the moment the materials aspects became very important, we left. Clearly with superconductors we were not in this game.

When we went to liquid crystals it was a little bit different. On the one side, there was great need of invention. Chemical invention was stimulated by the search for useful materials. In fact this case of liquid crystal was the first time I saw a molecule really built for a purpose. Bob Mayer, who was working with us in Orsay, had the beautiful idea that if you took a certain type of molecule which likes to make a tilted smectic phase, if you used a chiral molecule as a starting point, this tilted phase should be ferro-electric. And this idea came to him while queuing for lunch at the Orsay cafeteria! He talked to us, then he came back and he induced some chemists – Patrick Keller and others – to construct a molecule like this. A few months later we had the first liquid ferro-electric. This I really look on as a landmark.

Bernadette Bensaude-Vincent (BBV): So would you define the materials approach as the design of molecules for a specific purpose?

PGDG: Oui, I think that it is a clean description. There is a lot of wishful thinking where people claim that they do materials science. Often they construct objects and build molecules without knowing what to do with them.

HA: Working with chemists seems crucial for building molecules. Were there chemists in your Orsay group then at the Collège de France?

PGDG: Yes we did have chemists in the Orsay group Liébert, Strzelecki and Keller, three chemists. They did a lot, especially in polymerizing liquid crystals structures in order to get stable structures. They had their own lab. We had a cluster of seven laboratories on liquid crystals and they were one of the seven. At the Collège de France when I came we had a very similar situation. For one, we had Jean Billard who was working in close cooperation with a chemist at the Collège. And Jean Jacques, who was a chemist a great man who is dead now took one of his best chemist coworkers – Maya Dvolastsky – and he asked her to go and work in my lab. In fact she worked for twenty years with my group.

HA: And during the superconductor period?

PGDG: As I said during the superconductor period we were not materials inclined. We left the subject when it became materials science.

BBV: Do you think that it was the subject of liquid crystals that led you towards a materials approach or was it a more general trend in France in the late 1960s?

PGDG: A little later when we became interested in polymers. We were stimulated by the notion that you could get some useful product. This was a time, after the 68 movement, when we began to feel that we need to be useful. For the liquid crystal project, it was intermediate. We had the notion that these materials had to be useful but we did not think that it was our duty to invent systems. We were interacting with people at Thomson who were very close – one mile from us. I had a great admiration for the Thomson research lab because they had been very active in laser research. They had a very clever advisor Pierre Aigrain, but the French activity in liquid crystals activity was not very brilliant. Looking at this time from a distance I think that had it been ten years later, we would have taken dozens of patents. At the time, the push towards application was not very strong. (I admit that we would never have invented this classical display that we have in our watches because to me it would have looked too complicated. I would have been afraid of producing the twisted system in industrial conditions. But who knows?)

Then we went to polymers and many of us began to interact with industries. Around 1975 we really entered into an industrial network.

HA: Is there a continuity between superconductors, liquid crystals and polymers?

PGDG: For the liquid crystals, I think we have been lucky. The Russian school had a glorious past. They could have done an immense amount of work but they did not go far enough because they had a prejudice against chemistry and dirty materials. Because of that we could set up a French activity on liquid crystals without having Russian competition. It was a great luck.

HA: However you were not an advocate of dirty science?

PGDG: The tradition of superfluidity was very clean. The materials we were using were model materials with few defects. When Anderson used the word “dirty superconductors” he meant alloys. It is true that the physics of alloys has been very different for superconductors from the physics of pure metals. You can reduce the correlation length, you can control it by choosing the mean free path. There are many facets that become available when you accept to work with alloys. But in my mind, these alloys were perfect alloys without any precipitation or any complicated effect. They were ideal materials, although Anderson used the term dirty alloys (for provocative reasons).

BBV: Could you please clarify YOUR notion of dirty material?

PGDG: I don't use it often because in many cases there is a prejudice. My own distinction would be slightly different. It would be between universal and zoological. Let's take a different field like interfacial science: you can find universal features in this. You can construct general laws. The statics and dynamics of wetting are also pretty universal. But if you have a very specific problem such as making a polymer hydrophilic on its surface, then you enter into a certain amount of zoology. For instance to create a hydrophilic surface by a plasma treatment, this plasma treatment works in an unknown fashion with empirically chosen gases, under conditions that are not deeply understood. Details on a chemical surface are not universal and when you work for a practical purpose you better go into these details. Our attitude as physicists was to start from the universal features…with the hope that it would be useful for applications later.

HA: This is a physicist's perspective. But when dealing with polymer materials you had to extract some cleanness out of dirty stuff.

PGDG: It is true that there is a huge conceptual gap between semiconductors where you look for impurity fractions which are amazingly small and polymer physics where in all cases you will synthesize a polymer by a process which has some randomness. However you can build up universal laws despite the intrinsic distribution and complexity of these materials.

HA: What lead you to soft matter science?

PGDG: There is an amusing historical aspect. We had been working on superconductors when one day we had a beautiful seminar by Charles Sadron, one of the founders of polymer science in France. He started from polyethylene (that we suck when we suck milk from a bottle) and moved to considerations on DNA. He covered everything, in this wonderful talk. Our little group in Orsay was fascinated by his talk and we decided to go that way. Sadron's lab (then directed by Henri Benoît) was brilliant not only in science but also from the human aspect: they accepted us coming with our questions sometimes relevant and often stupid. They really established a co-operation with us. We worked for two or three years on polymers (it was roughly in 1966). We produced some little theoretical reflections on the dynamics of chains in solutions. But we didn't have an experimental lab with us in Orsay. This situation of hanging on theory exclusively, I did not like it. In 1968 or ‘69, we heard about liquid crystals by Georges Durand. He came back from the US and told us it was something for the future. We listened to him. So we suddenly shifted from polymers to liquid crystals and we worked on it for about 5 years. It was a happy period because within a few months when we crystallized the idea we got seven independent units cooperating on this project. There were chemistry, as I mentioned, nuclear resonance, defects (Friedel was very helpful because there was a tradition), optics, theory, crystallography. I may forget some of them but it came up to a bunch of six-seven groups working together in a happy way. Funding was easy. These groups were not nervous about their future; they were very open and willing to go into something like that. It was a great time to connect all these good people and just working together. The results were obvious. Within two or three years there was a French science on liquid crystals. There had been one fifty years before with Georges Friedel. But there had been a gap with only one group flying the flag energetically, the Chatelain group in Montpellier. They were lonely, however. They had a good education in liquid crystals but many tools that were obvious to us - such as inelastic light scattering, or nuclear magnetic resonance - were unknown to them. So to come back to our point, we in Paris could set up something very efficiently in a short time and one of our sources of pride was that it cost no extra money to the taxpayer. Because all the equipment was already there, there was no new costs.

BBV: You mentioned that you took advantage of the large apparatus in Orsay…

PGDG: Not big. It was not synchotron or reactors, no large machines.

HA: And neutron scattering?

PGDG: There was some neutron scattering on liquid crystals but it was minor. X rays yes; we used a lot of x rays especially when we came to the more zoological work with a long list of smectic phases which are more and more complex. But no large apparatus.

HA: When you began on polymers was there a lot of experimental data from neutron scattering?

PGDG: We came back to polymers after liquid crystals. I was at the Collège de France. We established a three-group collaboration with Strasbourg (Henri Benoît, a leading figure) and a group with Gérard Janninck at Saclay on neutron scattering. Here neutron scattering was very helpful to examine the conformation of one chain in a dense system where there are many other chains. If you have isotope labeling you can have this chain labeled and you look at this chain and describe the conformation of one particular chain. In that case, there was an old prediction by Paul Flory that this chain would behave like an ideal random walk – which is surprising for a strongly interactive system. Indeed the Janninck group proved that this was the case. So we had this cooperation, we were what we call in French la mouche du coche, a little fly stimulating the carriage but we had very little meat. Gradually however we got some. Francis Rondelez installed clever optical techniques. At the Collège at that time we had two types of activities. One was polymers the other one being surfactants. It was the time when young group leaders became advisors in industries. Christiane Taupin, who worked on surfactants, went to Levallois to head a group of Atochem. Francis (Rondelez) was an advisor to Elf, I was an advisor to Rhône-Poulenc. We worked more and more in close connection with industry. This was another happy time also based on cooperation. However it was a different cooperation, no longer a federation of little groups but a cooperation of large units, like Strasbourg. In Strasbourg they had a culture in light scattering and H. Benoît had constructed very detailed descriptions based on light scattering. Suddenly they were given the neutron scattering with isotopes providing information at a smaller scale (50 Ångströms instead of 5000). They were immensely happy with the neutron and Benoit wrote a book about neutrons and polymers.

HA: Nevertheless you spoke in critical words about big instruments.

PGDG: That was later. From the 1960s to 1985, I was a supporter of them because they had an educational aspect. This may be specific of European countries which have been delayed by the war. In the provinces, France had excellent abilities but no education. If you took young scientists from the lonely sites and brought them to Grenoble or to Saclay they learnt very fast in this intense research milieu using many concepts they had never heard about. They came back to their own labs and brought what they had learnt. So it was immensely useful. I think that the early generation of big machines has been excellent. At this moment I am less enthusiastic because the educational problem has been solved, fortunately. A student in a small city in France can have a good education in basic physics of condensed matter. From the point of view of discovery, the density of discoveries around big machines has dropped down fast. Let me take an example. Going back to the far past, in 1957 (the year when the Russians launched Sputnik) at the first international conference that I attended as an engineer at the CEA in Stockholm. It was about neutron scattering. I learnt two things from this meeting. I heard a talk by Harry Palevsky, a student of Fermi. He was an invited guest for six months in Stockholm. He had worked on the very small Stockholm reactor using energy selection methods which are very primitive – it was just a beryllium filter it does not provide a peak in energy, just a step. Using only that, he has been able to study the protons of helium. That was beautiful! It taught me in some sense that you could work with simple means without big machines. The second thing I learnt was the danger of theoretical gurus. There was a number of them at this meeting, in particular Walter Marshall and Roger Elliott from England. I was just a PhD student. I came and said to Roger Elliott that he wrote something wrong in a review article. He pushed me out although he was wrong. That taught me some caution with the old gurus.

HA: Let us come back to adhesion. It was a good example of a dirty problem at that time based on some science and on empirical rules.

PGDG: You are absolutely right. We entered into adhesion after spending some time on wetting which is more fundamental. I was struck by the great chemical successes achieved in adhesion. The example that I often quote is anaerobic adhesives. These are systems that you want to reticulate, to polymerize once they are in a proper position between two walls but you don't want them to react stupidly in other situations. In that case, chemists were able to have a polymerization induced only in the presence of certain metal surfaces like copper. That is chemical invention. My impression is that chemistry has been the leader in this field. We physicists, and the people from mechanics, we were in a more modest position. People from mechanics brought measurements. To define adhesion properly instead of measuring the force between two pieces, Griffith and others established that you have to measure the separation energy per unit area. So people from mechanics provided 1) the measurement techniques like the cantilever technique and 2) new concepts. Thus chemistry ranks one, mechanics two. Physics comes only as number three. So in adhesion meetings you could hear these nice theoretical talks – not easy theory indeed – but very nice. At the end of such talks somebody raised his hand and asked: “what does it tell me about this particular adhesive where I found that when I modify my molecule by putting this methyl group in the sixth position I get a much better adhesion than if I put it in the fourth position”? So that was a kind of Babel Tower. Our modest aim was to try to build up a common language. We helped a little bit in two respects. One is the question of very soft adhesive materials where dissipation inside the adhesive is what makes a material good. We could help because it was close to concepts we had met in polymer science. The other question concerns little polymer chains that intertwine. Liliane Léger has been working on it. We thus had, let's say, two years full contribution but it was very modest. It did not clarify the science of adhesion. But it helped create a number of teams in France. If you look at the situation I would say we have

1)      a classical lab in Mulhouse where modern physics was introduced by Günther Reiter

2)      Costantino Creton here in PC (the Ecole supérieure de physique et de chimie industrielles)

3)      Liliane Léger on polymer systems at the Collège de France

4)      a small group with M. Shanahan in Corbeil.

HA: Apparently it took time for you to convince them to work on such a subject.

PGDG: Absolutely right. My dream would have been to set up a sort of adhesion science center in the Paris area. Ultimately I did not manage to do it. There are various scattered researches but no unity although it is not too bad.

BBV: Did you work on adhesion because you had industrial contracts or was it your own initiative.

PGDG: I think it was our own initiative, although I may be wrong because it is very difficult to trace the origin of a project. We had no program with 3M, the great master in industrial adhesion. Rhône-Poulenc had some related problems but they don't sell adhesives as such. Latex is special: it is not a real adhesive. We heard about adhesives but it was not something important for them.

HA: Gilbert Schorsch from Rhône-Poulenc?

PGDG: He was more concerned with new materials, organo-mineral materials. Later they turned to adhesives. I don't remember well. I think that the wetting problem, dealing with interfaces led us to move to strongly interacting systems. But we should be very modest. Take for instance a standard adhesive material like the epoxy-glue that you buy in a supermarket. Frankly, I don't understand the way it works.

We are still working on adhesives. If you look at this blackboard here you'll see that recently we have been concerned with cellular adhesion. We have a professor in medicine in Marseilles, Pierre Bongrand, a former student in a solid-state graduate school here in Paris, who brought a number of key measurements in adhesion. The notion is of a cell with a few sticky molecules at its surface but they are very small, very dilute. When the cell comes in front of another one, all the sticky molecules move to the contact region and build up bridges there.

Ten years ago Bongrand and others understood the statics of that process and what the separation energy is. It is not at all what stupid people like me would have believed. When you begin to separate you do not have to cut a bond because all the stickers just go to a smaller surface but they don't disrupt their bonds. So the adhesion energy is just fighting against the osmotic pressure. People like these established deep ideas about the statics. While I had to give a course I realized that there was a cascade of problems concerning the dynamics and I started thinking about them. So we are still on adhesion.

HA. How do you see the links between biology and materials research?

PGDG: I have been very critical about biophysics. For instance, physicists had in mind that they could do a lot of biophyscis on cellular adhesive molecules by establishing the 3-dimensional structure of these proteins. It is helpful. However, it is not a very exciting program because the biologists are so clever that they immediately sequenced these proteins. They realized some parentages between the sequences, grouped them into families and could identify the function of the various pieces without big instrument physics. The interesting problems – how does it work in a tumor situation, or how do I influence this process, how do I stimulate them – are not in biophysics. Biophysics is doing only the details, not addressing the big question. That is why I have been so critical of this community who jumped into biophysics at one stage. Fortunately I was partly wrong. There are good examples around here: at the Institut Curie Center with Jacques Prost, they really have a wonderful activity. For instance, they have a universal theory of molecular motors. That is a real success of biophysics. There are facets of biophysics that I respect very much but there are still old facets that I would call more engineering than science.

HA: Let's talk about the artificial muscle.

PGDG: The subject was started by a giant in polymer science, Katchalski. Polymer physics started with Kuhn in the 1940s. Ten years later Katchalski, a former student of Kuhn, said: “if we understand rubber, maybe we can devise a rubber or a gel where a chemical agent changes properties, transforming chemical energy into mechanical energy”. That was a beautiful idea. Katchalski did very sophisticated work with very simple means. He had very few materials available in Israel at the time: he used methylacrylate recuperated from the cockpit of World War II aircrafts. He did a wonderful job. The materials he produced demonstrated the principles but they could not have any practical application because of slow response and fatigue problems. This historical contribution raised an interesting challenge.

We started as a small thing. With gels, the response time was very bad. Then we tried liquid crystals systems with no solvent. You just changed the temperature to switch the conformation. That was tempting. So we have a project at the Institut Curie which requires delicate chemical synthesis. Another project was launched by a Japanese team in Osaka. They are electrochemists and they used a membrane made of a popular material for other purposes in large-scale electrochemistry. This membrane is called a Nafion; with this Nafion they were able to achieve systems that under moderate voltage – a few volts – distort and then command actions. The response time was around one second. I am full of admiration: not only did they build up the material with the correct (large-area) electrodes but they also understood the dynamics of the process. The field is very attractive (but our contribution is very, very small).

HA: Would you say that artificial muscle is a bio-inspired material?

PGDG: It is not really bio-inspired. It is based on polymer science and has nothing to do with an actual muscle. But I am fully convinced that bio-inspired materials will become more and more important. I was very impressed by the German team that found what are the peptides at work in making the shell of diatoms. Using this sort of results in the future is very tempting.

BBV: Did teaching play a part in your continuous shifts from one project to another one?

PGDG: Ah oui, teaching played an important role. This figure on the blackboard on cellular adhesion really came from the fact that Françoise Brochard was having a course on soft adhesives for industrial people. When she asked me to talk about cellular adhesion I just realized that I could not teach it because I did not really understand the process involved. So it was an excellent push. Teaching is very helpful for theorists because we are often trapped in formal models. Mathematical writing does not give any idea of the real thing. We have to re-digest and transform the mathematical statements into a few simple sketches without any calculation. Teaching is good for going in this direction.

HA: What about your experience as a Director of an engineering school?

PGDG: I have tried to keep scientific contact with the labs, on gels, on separation techniques and some other cases. I try to keep this place aware of new fields and to keep good contacts with local people. I sometimes missed the point because of too many duties but right now I am very happy. We have new young professors such as Jérome Bibette working on emulsions, Ludwig Leibler on polymers, Jérome Lesueur on transport in superconductors and it is very stimulating to talk with them. The person you really want to direct such a place is somebody who is able to talk to everyone. I would be scared to have separate departments for physics, chemistry and biology.

BBV: Do you think that throughout your career you crossed disciplinary boundaries, or are you still a physicist but able to talk to other disciplines?

PGDG: I tend to see it more as a process of learning. For instance when we entered the field of polymers we were like students. As I said, we made many mistakes. Our lives have been a cascade of student lives. At least this has been my feeling. For the theorists it is easier to move, they can switch more easily than experimentalists. But some experimentalists did switch: Etienne Guyon for example moved from superconductors to liquid crystals and granular matter. He is an interesting case.

HA: Do you intend to pursue your recent interest in glass?

PGDG: The literature is difficult to grasp. We look at a certain sector, mainly on structural glasses, with problems in real space, numerical local space features.

This page was last updated on 29 May 2002 by Arne Hessenbruch.