Sciences : histoire orale
Accueil du site > Individus > WHITESIDES George, 2002-01-28


WHITESIDES George, 2002-01-28

Enregistrer au format PDF

George Whitesides

Mallinckrodt Professor of Chemistry, Harvard University

George Whitesides’ group in Harvard’s Department of Chemistry and Chemical Biology works on many topics : Biomimetics, Biosurface Chemistry, Complexity / Emergence, Materials Science, MEMS, Microfluidics, Micro- / Nanofabrication, Photonics, Polyvalency, Self-Assembly, Surface Science, and Tools for Biophysics and Biology.

Pour citer l’entretien :

« Entretien avec George Whitesides », par Bernadette Bensaude-Vincent et Arne Hessenbruch, 28 janvier 2002 Sciences : histoire orale, https://sho.spip.espci.fr/spip.php?article130.
— 

Entretien avec George Whiteside, par Bernadette Bensaude-Vincent et Arne Hessenbruch, 28 janvier 2002 : on history of organic chemistry and materials research and on the purpose of history of science.

Lieu : Prof. Whitesides’ office at Harvard University

Support : enregistrement sur cassette

Transcription : Bernadette Bensaude-Vincent et Arne Heseenbruch

Edition en ligne : Sophie Jourdin

BERNADETTE BENSAUDE-VINCENT (BBV) : Do you see any changes that you would like to mention in your field over the past 20 years.

GEORGE WHITESIDES (GW) : 20 years ago there was not any materials science in my field, which is organic chemistry. People made polymers, but they didn’t really think in terms of materials, and all of this stuff of electronics and optics and unusual functions and biocompatibility and things like that were - some of them were not even concepts and some were not considered materials science because nobody thought about materials science. Materials Science in the last 20 years has come from something that was done with airplanes to be a real part of chemistry - that is really interesting I think.

BBV : Have chemists been involved from the very beginning or did they jump on the bandwagon later ?

GW : Well, no, I guess my sense would be that chemists were involved from the very beginning in the sense that they made materials. Because after all most materials in the world are actually made by chemists.

We have metals, but ceramics often come out of chemical backgrounds, and all of polymers come out of chemistry but they were thought of as being fibers or coatings or something else and the idea of thinking about them as materials - it’s a re-naming but it was helpful for the field to think about how to organize itself intellectually among other things, and then what was a little bit of a change was that at the beginning the financial support was primarily from the DoD for its own purposes which was primarily metals and ceramics for airplanes and armor and hot-sections in engines and things of this kind. It is recently that this has changed. Chemistry and chemical thinking is an increasingly important part of materials science which means there is now some money which means it is worth the effort for chemists to think about that as a community they might want to be involved in. But chemistry has gone from an area dominated by the notion of making natural products to an area in which probably the two interesting areas are bio - and most people would say this is the biggest deal in chemistry at this point - and materials. Synthesis becomes an accessory technology which enables one to make these materials that you might want for either biological applications or materials applications.

BBV : And what about the electronic applications for chemistry ? Is there any future in that ? You have been working a lot in that.

GW : I think that is going to be one of the big deals in electronics. The argument there is that there is a certain specific way of thinking about electronics which is silicon and semi-conductors and that works fine. But my prediction is that there is going to be a new class of devices that is going to be perhaps slower but much cheaper, so that the ratio of benefit to cost goes up but primarily because the denominator goes down rather than because the numerator goes up. The notion with these and with optical photonics systems and with organic LEDs is low cost, ease of fabrication, the ability to integrate a bunch of different kinds of functions in the same thing whereas if it is all polymers you don’t have to worry about compatibility between the 500-degree processing step and an ion-beam step or whatever it might be. That is organic chemistry. There are two components : one is the laser and the second is the coating. Optical recording now dominates a certain segment of that industry and it is entirely because what chemistry has made possible. CDs are the same thing. These are optical CDs but the older CDs were the same thing. I think chemistry has found opportunities in thinking about the collective properties of matter as opposed to thinking of the properties of individual molecules - it is a little bit of a revelation but it opens all kinds of doors.

ARNE HESENBRUCH (AH) : Shall we try to get at the history of the last 20 years ? What in your opinion has been the most important changes in the last 20 years ? You have been in the field for that long.

GW : It is more like 20 years. I really began to think seriously about materials science when I came to Harvard. At MIT I was primarily concerned with inorganic chemistry and catalysis. So I came here and became involved, but what made a big difference from our point of view was the development of self-assembling monolayers. Self-assembling monolayers were particularly interesting - and I think actually very interesting for many chemists because they made the point that you could design molecules and thereby control macroscopic properties. You make a kind of molecule and a drop of water beads on the surface and you make another kind of molecule and a drop of water spreads on the surface. So that idea of being able to engineer macroscopic properties based on molecular-scale synthesis was a pretty interesting idea and fairly new at the time.

AH : When ?

GW : That happened ... the first identification of SAMs per se was early ’80s and forget exactly when, whether it was ’83 or ’84, and then the real understanding that you could put them together was done by Ralph Nuzzo - he did the first identification of these as structurally ordered systems. And then we did a lot of the work that connected them with materials ; we were actually working on it at the same time and we started publishing papers in the period of 88-89. So it’s really only been about a decade. Now polymers have been around for a long time and we have been working on polymer surface chemistry in the late ’70s. But we didn’t really call it materials, we called it polymers and it was very hard to do because you couldn’t really get a grip on it. You’d start with a piece of polyethylene and we used to put in chromic acid solution and that burned the surface and introduced a bunch of carboxylic acid groups which we could use for various purposes but it was very hard to characterize these things and the surfaces were pitted as a consequence, so we never really had a good grip on the density of surface functionality, the roughness of the surface or any of those things.

AH : What did you do to characterize ?

GW : We used XPS and we developed a whole series of techniques like contact-angle titration - which is an idiot idea but the notion that you make the drop more and more basic and you see if it spreads more and that tells you if there is an ionizable group. This is the kind of thing a chemist thinks of first but not a materials scientist with a metallurgy background because they don’t know that carboxylic acid groups exist. Self-assembled monolayers still are, I think, the best of all the materials systems - and this is almost independent of field : ceramic, metal or anything else. If you want something which you can control in exquisite detail it is this. If you want to understand how you go from atomic and molecular level structure to macroscopic property within the range of things that it can do, nothing beats SAMs.

AH : And you knew this in the early ’80s already ?

GW : No, what you knew in the early ’80s was that there were ordered structures. Strictly speaking, well before that, the electrochemical guys had known that organic silver compounds adsorbed on metal electrodes. So that was well understood and they’d been used empirically for a bunch of things having to do with adhesion so there was, as in everything in science if you look backwards you find, lots of precursor phenomena. And then Ralph Nuzzo and a guy named Dave O’Lara(?), who were at Bell Labs at that point got involved in this and I think their initial interest (I actually never asked Ralph why he got into but I think it was partially because they needed ways of doing coatings of metals and - anyway they were good spectroscopists - and Ralph looked at these structures and recognized from the infrared spectra that they were highly ordered so that the idea was that it was not just going on the surface as scrambled eggs but that it was crystalline. And then we did physical organic chemistry. We took that platform starting in the early ’80s and put functional groups on the terminal positions so that after the thyols formed mono-layers, you just saw a monolayer of the terminal functional group, and the hypothesis was that this would control wetting. And it did. Then the process from there was that in the early ’90s we had the idea that one might be able to make patterns with these self-assembled monolayers and we had several approaches to this - the earliest one was actually an experiment I did with my own two hands - which was to take a piece of a gold dish (gold on silicon), put on a hydrophilic thyol so that water spread on it. We took a razor blade and put some scratches on it and redipped it in a hydrophobic thyol, so one ended up with scratches about a 100 microns wide they were hydrophobic in this sea of stuff that was hydrophilic then put drops of water on it and the interesting fact was that the water would not cross the lines. Had the characterization changed by then ? By that point the characterization was in pretty good shape, the combination of reflection infra-red and XPS did a lot of it. I forgot when people first saw STM images but it was ...

AH : The Nobel Prize, DI began selling in 1989.

GW : Right, and they were hard to see at the beginning because they require the tip to be far away, so it was not the first thing that was done, but people began to see crystalline rays (?). We had done a fair amount of work in the mid-80s on transmission electron crystallography. We’d taken awfully thin gold foils and run electron beams through them and seen diffraction from the organic mono-layers I think those were the first real structural indicators other than the infra-red spectras that they were united. So anyway, we started the business of stamping. I think the first paper was 92, 93, something like that. That became the basis for micro-contact printing and then for soft lithography and that brought an entire new community into it who are now on the other side of the electronic guys who are looking for new ways of getting around diffraction limits and looking for a different economics. And it has not yet made an enormous difference in microelectronics although it may well in this next stage of all-organic electronics. But soft lithography was probably the most important thing in moving the micro-fluidics community away from silicon to polymers. Now, in the long run whether the microfluidic systems we use for genomics are based on polystyrene or PDMS or what, remains to be seen - it will be whatever is cheapest. The basic idea that you did not need silicon, and in fact didn’t want silicon or glass for microfluidic systems : all you wanted was that it was cheap, transparent, and easy to work with, and that is polymers. And now you put those kinds of things together with the idea of conducting or photoactive polymers and ever-increasing resolution of these printing and stamping techniques and the economics which are just phenomenally better than traditional methods when you don’t need the kind of precision that the photolithographic methods bring, and you find all of a sudden that there is a new approach to making micro- and nano-structures. It is an interesting evolution. What about the biological ? Well, the biological has also been a big deal and the issue there is interesting on a couple of levels : one of them is that biology since the 60s, which is when practical sequencing first really started, has been profoundly reductionist. The whole notion is that you get the gene sequence and everything comes from that by some unknown process. That is not true of course. People now understand that it is not clear what one does with the information of the genome : which will be useful, how you combine that with proteomics, which will be useful in cell biology, which will be useful in knock-outs, which will be useful in the ... and so on. These are all interesting questions that we don’t know the answers to. Biology is pretty complicated. In developing biology, as in any field, you need new tools. And new tools are microfluidic devices for genomics, new ways of manipulating cells which have to be in fluids.

AH : Could you put some dates on those new tools ?

GW : They were done in silicon in, maybe, the 70s. The guy who did this was named Andreas Manz who is now at Imperial College in London. At that time he was at Sandor ... Ciba-Geigy ... I forget.

AH : Basel, anyway.

GW : He was in Basel - I think he was in Ciba-Geigy. Of course it was an internal project so none of it was published at that point. He had two post-docs, one named Mike Ramsay, the other named Jed Harrison who were both migrant ; Mike to Oak Ridge and Jed to Edmonton, Alberta, where they started to make microfluidic systems. And that is really where microfluidics started. Other people, Roger Howe, did have an important contribution. In fact, Andreas made some of his first systems out of PDMS - unpublished stuff - he recognized the virtues of this nice elastomeric polymer even back then. But he couldn’t publish it and I think the company wasn’t very interested, so it sort of languished for a period of time and then this genomics came along and there was this need to find better ways to crank out data large quantities of data. Then there was a real need for microfluidic systems that were parallel with which you could generate lots of information. That was an enthusiasm that began to build in the 80s, late 80s, and people had been making devices using silicon microfab technology and it was clear it was just too cumbersome - it took too long ; they didn’t have any control over the surfaces and then we among others made the point that if you use self-assembled monolayers you could get exquisite control over the interface between the materials world and the biological world. We have done a lot of this in collaboration with a guy named Don Ingber here at the medical school. This is in the late 80s, actually the most interesting part of this began came fairly rapidly after we began contact printing or that is when you saw some really neat things. So this was, I would say, 1994 or 1993, I forget when the research was done, but in that period of time. So that was one theme in this : that you could make the inorganic materials of gold films, or whatever they were supporting, you could make them compatible with biology using self-assembled mono-layers. But at the same time other people had begun to tinker with different ways of making micro-fluidic systems. I don’t know that I can say what all the threads were, but certainly one of the threads was a discovery that we made that instead of using photolithography and clean rooms and all the rest of that you could do the design, print them out on a printer at scales that were perfectly useful for this kind of work and then you could design quite sophisticated systems. Other people were designing quite sophisticated systems at the same time using conventional lithographic methods. We made a point that you could do it very easily using these soft lithographic methods and that provided quite a stimulus, not to the generation of the final systems necessarily, but to the process of invention because lots of people used those rapid prototypic methods for trying new kinds of devices. This has also been an area where the industrial groups, particularly start-ups, have had a very big influence on things. Probably the first company that was a serious effort in this was Caliper. Caliper was a Larry Box(?) special and it’s doing okay as a company. It is focused now on micro-fluidic systems for high-throughput screening for the pharmaceutical industry - it was one of the first to do that. There was a fellow named Lee Hood(?) who was named Hunkerpillar(?) - I think a bioengineer - developed a series of devices for sequencing of DNA and proteins which were commercialized by a company called Applied Biosystems which was one of the early darlings of the biotechnology industry, and those were not intentionally micro-fabricated - they just turned out so because you needed lot of them and small. They ended up making nice small systems - not very small but small systems. And then there was another thread to this which was the whole notion of high-throughput screening which had a number of connections that tied together. One of them was just the need in industry to do lots and lots of experiments in parallel the acceptance of 96 ??? plate formats which was the first of these. And then the idea of parallel synthesis, combinatorial synthesis, parallel synthesis which came partly from Bruce Merrifield – he got the Nobel Prize for solid state synthesis of peptides - and partly from a bunch of people including, at least I give credit to, a man named Mario Geysen as being the first person to do combinatorial synthesis, parallel synthesis, on pins. They weren’t terribly small but he had that notion. Out of that came the whole idea of very large-scale parallelism in synthesis which had a requirement for making things small which led to new requirements for small systems. It is an interesting area, the bio-area, both because it has had an important impact on genomics and drug discovery and also because it provides a bridge between these two areas of biology and materials science. This is where they come into contact.

AH : You have put together elements of techniques and ideas. I presume that funding structures have had impact on the history also ?

GW : Yes. Most of the funding for materials science in the US has come from DARPA. DARPA also got interested in the biological stuff in the mid-80s, maybe a little bit later than that. Their driving concern was with biological weapons - biological defense. The United States does not have an offensive program but it was clear that there was a potential problem there and then in the late 80s there was a Russian defector and then in the early 90s another, Ken Alabakoff(?), who spent a lot of time talking about what the Russian program had done. It was clear that the Russians had had a very very large program in biological weapons and it really bothered people. So there was a growing effort to think about how to build sensors and analytical systems, how does one put together a defensive program in this. That ended up being an important bridge between the materials community (which was the DARPA community) and the biological world. One of the things that DARPA does very well as that when it decides to go into an area to put in enough money for a sufficient period of time that it is actually worth people’s effort to go and learn how to do that because they can run programs that make a difference. They get, I think, a lot of credit for building the biological component for micro-fluidics and then of course a very large fraction of soft lithography was funded by DARPA and their interest in that was largely new methods of circumventing the natural limits in photolithography for very high-density electronics. It has not in fact changed that - I mean these techniques are not used for high density electronics - but I think they are going to be the basis for low density low-cost electronics, the organic stuff. I think this whole area is one that DARPA gets the major credit for. Had DARPA not put the money in it would not have happened or it wouldn’t have happened the way that it has. NIH has been very resistant to doing anything in this because NIH only does science, it doesn’t develop tools. I think it is an enormous strategic mistake on their part but they have basically been parasitic on activities funded by other parts of the government or funded in industry, in start-ups, for the tools that they use and the biological community uses. Then of course there has been some work in the NSF and DoE but not big programs.

BBV : So do you see the contribution of your group as providing new tools for materials science ?

GW : The biggest contribution is to make the connection between physical organic chemistry and materials science - that is the idea that you can design molecules, or design matter by synthesis at the molecular level, and thereby influence macroscopic materials properties. The second is to put together a series of tools, particularly self-assembled monolayers and soft lithography and illustrations of the uses of these in microfluidics and particularly in bio which have caught people’s attention. This is an interesting strategic decision : we decided at the very beginning that we were going to proselytize for the area, so we made a really big effort to have anyone who was interested. We invited them to the lab and tried to teach them everything that we knew. It is one of those interesting philosophical issues in science : when you have something that is new and interesting you either try to hold it close or you try to spread it as widely as possible, and we really made a big effort to spread it because I thought it was going to be very useful. Quite a large number of people would come for two-three days or a week over the course of time and then these people have gone out and started programs and then of course the graduate students come and learn how to do these things and they go off and start programs. So, the dissemination has been quite rapid, primarily because it is very easy to do. That is another philosophical issue : I like to do research that is really easy to do because other people can do it. If you want to have an impact on the community at large it is easier if you keep the bar low to people getting into the area rather than making the bar high. So, minimal capital investment.

BBV : And how much time do students need when they come to be trained in the techniques of self-assembly ?

GW : Two days.

BBV : Two days ? It is kitchen experiment !

GW : It is kitchen experiment. You take some wafer and you dip it and you pull down the stamp(?). That is not complicated. And stamping is literally no more difficult than putting a rubber solution on a rubber stamp and putting it down by hand and pulling it off and it’s done. It obviously gets to be more complicated than that but you’ve got the basics at that point. There is a little issue there, which is : you say this is trivial, and in a sense it is trivial, but it is trivial because some very smart students spent a lot of time understanding these systems and figuring out which ones worked so well that they could not be made to fail. That is the criterion. The simplicity at the end is the result of a lot of hard thought along the way not just by us but by the whole group of people who was interested in self-assembly and related areas. The second thing is the tools and the third is the notion of self-assembly. Generally, organic chemistry has been largely concerned with making molecules by putting together covalent bonds and self-assembly is how you make materials through non-covalent interactions. So it is a switch in emphasis. We are certainly not the only ones who’ve worked on self-assembly but we have probably been more emphatic in making the case that it is a technique that can be used in materials science than other people who have been concerned with making structures in solution or how proteins fold or other things of that kind.

BBV : It is striking that other groups doing self-assembly have had biomimetic inspiration whereas in your group it was not a leading ...

GW : Well, we do biomimetic stuff. If you talk to some of the people around, Mila Boncheva, who I think with a little luck you will meet, has made this wonderful device. You basically have a string of something and you put this in suspension, it folds up and one part of it becomes a high-frequency ring oscillator and the other part becomes a shift register. It’s quite intentionally patterned on the idea of protein folding so it is biomimetic in that sense. But the major issue is the generic problem, which is to understand molecular recognition in biology because if you want a single process in biology that is important it is the recognition of a ligand by an enzyme-active site, or whatever - that is the reductionist end of the working part of biology. There has been lots of work on that - and then self-assembly came in, as a result of two things : one was to try to model biological recognition and interestingly this has never really worked. There is an enormous amount of molecular recognition that has gone on in methylene chloride solution, but of course water doesn’t go on in methylene chloride, so in organic solvents a lot of good work has gone on which has led to systems that are pretty efficient in molecular recognition. But noone has come really close to duplicating the efficiency of biological molecular recognition in water because we have to learn how to handle the hydrophobic effect and we don’t know how to do that yet. So that is one theme. The second theme was the stuff that Jean-Marie Lehn, Cram, and Pedersen from DuPont got the Nobel Prize for, which was just understanding the crown ethers of the right size would wrap themselves around a sodium ion. That was not really molecular recognition, but it was self-assembly ... well ... Jean-Marie calls this supermolecular chemistry, so it is chemistry beyond molecules. It was an important first step in making the case that it isn’t just making covalent bonds that’s important but these events in which things recognize one another and get together without making covalent bonds are also important. So these two things came together in self-assembly. One virtue of this place is that students are very good and very independent, and the groups are big so that we can have biologists, chemical engineers, physicists, organic synthetic chemists, and materials scientists. That makes it much easier to think about putting biology and materials science together, or organic synthesis and microelectronics ; you know x and y, where x and y come from very different directions. It is a lot harder to do that in smaller groups.

AH : Do you think there is any purpose in doing history of science ?

GW : Oh yes. Understanding what this funny activity is of creation. Hang on, I’ll be right back - this last question is important ...

History of science. If you look at history of science you can see where interesting things came from. Now, the question is, can one learn something from the activity, and I think the answer is yes. The question that comes up in this area is, is it sufficient just to put different research areas together in order for good things to happen, or do you have to have a brilliant conception ? No, I don’t think in this case one needed a brilliant conception. It was putting things together - with some luck - self-assembled monolayers were an experimental observation, they were not a theoretical prediction confirmed by experiment. It still amazes me that soft lithography works as well as it did. In any event, one can ask where ideas came from, and from that point of view the history of science is an enormous contribution. The question which is never clear is, to what extent are there lessons that you can learn from one success or one failure that you can carry over into other areas ; and in particular : what is the mechanism by which whatever lessons you learn are transmitted to the people who are the ones who in principle would benefit most, who are the people just beginning their careers, but who are in general the least interested in listening because they are eager to get on with it. They have their own usually not very well-formed theories as to how to do things and it is based more on their own particular instinct. A lot of their instinct, in fact, turns out to be replication of the experience in graduate school. That is probably the least perceptible but longest lasting lesson of the graduate career, because people tend to do their research the way they learnt to do research which was the way their research director did research. That stylistic issue comes across. In a sense, one of the great sadnesses of science is to look at people at the age of 50 and realize they have spent their entire career working on their theses with minor variations on themes. It is not a happy outcome. What they learnt was, don’t step too far out of this particular circle ; just sit there and fret away with this stuff. So, I think there is enormous opportunity for the history of science to be useful, and even predictive if one could figure out what to do with the information. As we made a big effort to teach people how to use soft lithography, I think all scientific enterprises, particularly with fields that move rapidly, which is what science does : it is always looking forward and almost never backward, one needs some clever way of figuring out what the short-form lessons are from what you do, what you have learnt. But I don’t think the user will study things. A book which I think is a very interesting example of history of technology which has made a big difference is The Innovator’s Dilemma. Have you read The Innovator’s Dilemma ?

AH : I haven’t read it.

GW : You should read The Innovator’s Dilemma. It is by a guy named Christensen, who is somewhere here at Harvard or MIT. It’s history - history meaning business school history, but it’s close enough - and the question is : look around and you find big technology changes, and how do companies react to that ? And the answer is in general : incredibly badly. And this book has really caused people to think about how they run their technology programs (in companies) because the failures are so stark that as a big company person it really does not give you a lot of confidence that whatever you are doing is going to survive. One of the lessons in this is this : the mantra in the United States is listen to the voice of the customer ; the customer will tell you what you do. And one of Christiansen’s lessons, loosely re-phrased is that if the field is technologically changing the thing that will kill you is to listen to the voice of the customer, because the customer never wants change - the customer has people trained in some technique, they’ve got capital investment and so on and so forth : the last thing in the world they want is change. So you listen to your very conservative customer and some little company who wants to take your customer away or create new customers, is going to zip around on the side. Now the interesting issue for you all is, you are doing all this stuff and my guess is that websites may be interesting for other historians of science but I don’t think it will make any difference to scientists. So what you need is a nice short book which summarizes all of this in terms of lessons. What are the lessons that you can draw from this ? What has succeeded in what circumstances and why ? And what are the risks, and what has failed ? People hate failing. So look around and what are the failures, and what did they not do that prevented them from competing effectively ? I think if you think about history of science rather as the business school thinks about its analyses in terms of something that’s branded in terms of suggestions for the next generation or the current generation of operators rather than sort of a passive observation it might have more impact. I don’t know whether you care whether it has more impact or not.

BBV : We care about our impact.

AH : I care a lot.

BBV : Thank you very much !

Fin de l’enregistrement


haut de page

accueil du site


Haut de page | Accueil du site | Table générale

Contrat Creative Commons

Sciences : histoire orale de https://sho.spip.espci.fr/ est mis à disposition selon les termes de la licence Creative Commons Paternité - Pas d’Utilisation Commerciale - Pas de Modification 3.0 non transcrit.