The Science of Things: Unanswered Scientific Questions and Unquestioned Scientific Answers in Materials Research and Development

MRS BULLETIN/SEPTEMBER 2000 59 I believe Groucho Marx pointed out that the problem with prediction is that it is about the future, and we do not know the future. But if it is unprofitable to attempt to predict what will be known in the future, I see no such objection to emphasizing what is not known yet, but which ought to be known soon. Indeed, such an enterprise has two parts. That splendid scientific cartoonist, Sidney Harris, has a lovely drawing showing the doorway of an unidentified research institute, and in front are two signs pointing opposite ways. One says “UNANSWERED QUESTIONS,” the other, “UNQUESTIONED ANSWERS.” Sometimes, the most impressive advances come from questioning unquestioned answers. A good example of an unquestioned answer that eventually was critically reexamined is the dogma that a new phase must form from an old one by the process of nucleation, a discontinuous process in a metastable starting phase with the involvement of sharp interfaces. That model was worked out over a period of years (from roughly 1910, starting with Einstein, via several other German physicists, to Charles Frank and David Turnbull in the 1950s). Then, at last in 1956, Mats Hillert showed that a new phase could form by a continuous process in an unstable starting phase, and John Cahn and John Hilliard, a few years later, took Hillert’s notion further by working out the theory of a diffuse interface that gradually thickens as the unstable parent phase decomposes continuously into regions of diverging compositions. This completely new view has been immensely influential in modern materials science and the Cahn/Hilliard papers in particular are among the most frequently cited in our field. Another example of an unquestioned answer was the organic chemists’ dogma that a compound must have a unique, well-defined molecular weight and melting temperature, otherwise it could not really be a compound. Polymers do not fit that dogma, so for several decades chemists insisted that polymers were, in effect, stuck-together versions of small molecules. Until that conviction was overcome, polymer science could not properly advance. That unquestioned answer delayed the progress of polymer science by at least 20 years, roughly from 1910 until 1930. The way things have been going in recent years, I would guess that the field in which the most unquestioned answers will walk the plank is the processing of materials. Modern processing can be said to have started with the invention of splatquenching in 1960 by Pol Duwez at Caltech. He wanted to find a way of enhancing effective quenching rates of alloys, with a view to examining metastable phases, and found to his surprise that quenching from the melt was far more effective than quenching a solid sheet or wire. That insight was then followed through to its logical conclusion, and entirely new states of matter—metallic glasses in particular— were discovered and studied. At about the same time at General Electric (GE), the conviction that sintered ceramics must always be opaque because of unavoidable residual pores was overthrown when the link between grain growth and pore removal came to be understood; pore-free ceramic could at last be manufactured, and unexpectedly, it was exploited for lamp envelopes. More recently, the chemistry of self-assembled materials, in which regular arrays are formed either spontaneously or by the intervention of a template, has advanced to the point where a new materials encyclopedia, now being prepared, has an editor wholly focused just on self-assembled materials. One of the more remarkable developments in selfassembly is the discovery and systematization of colloidal “crystals” made from suitably coated polymeric spheres that, against all expectations, attract each other enough to form regular assemblages. (That research started from an examination of natural opal.) Colloidal crystals made their bow just as photonics (optical circuitry) reached the stage when crystals with a periodicity similar to the wavelength of light were needed to act as a kind of diffracting filter. This demonstrates that, in some mysterious way, need and solution attract each other from distinct regions of the scientific spectrum. Another form of processing that led to a wholly unexpected matching of need and solution, and which I suspect may prove to be an archetype of what the new century holds in store, was the exploitation of nuclear tracks created by energetic particles (fission fragments) passing through an insulating solid and revealed by etching. This was research done by Buford Price, Robert Walker, and Robert Fleischer at GE from 1961 on, and it was at first exploited for such purposes as age-dating geological and archaeological materials. Then, one day, the investigators discovered by chance that cancer researchers in New York needed an ultrafilter to remove viruses from fluid suspension. It transpired that thin plastic foils, when irradiated with fission fragments from uranium, could be etched to produce a uniform population of micronsized holes ideal for that medical application, and filters began to be manufactured commercially. Since then, with the expiry of patents, half a dozen manufacturers have created a multimillion-dollar industry for such medical filters. Another example of a chance discovery that led to major developments was Donald Stookey’s discovery, at Corning Glass, of FOTOCERAM, a light-sensitive glass that generates crystallization nuclei only when exposed to light, and can then be etched away where it has crystallized, but not where it is still glassy. Incredibly complicated arrays of holes and slits can be fabricated in glass or (by further development) in glass-ceramics by this approach. I am inclined to predict that the principle of photomachining, exploited in glass-ceramics and also by means of photoresist in the manufacture of integrated circuits, will ascend to new levels of subtlety in the years ahead. These examples of colloidal crystals, nuclear tracks, and photomachinable glass-ceramics pinpoint one of the issues for the new century. As we have seen, problems and solutions can come togeth“If an experiment works, something has gone wrong.” “You can never tell which way the train went by looking at the tracks.” “All’s well that ends.” —From a poster presenting Paul Dickson’s “Murphy’s Laws on Technology” The Science of Things: Unanswered Scientific Questions and Unquestioned Scientific Answers in Materials Research and Development