Taking the fuzziness out of quasicrystals.

Taking the fuzziness out of quasicrystals

By striving for perfection, scientists have gained a clearer picture of how certain types of quasicrystalline materials are put together. X-ray scattering experiments demonstrate that these unusual materials appear to have a fivefold icosahedral symmetry, a pattern that fails to fit conventional crystallographic rules. In the past, however, the fuzziness of the X-ray results left room for a number of different theories as to how such crystals are organized internally. Now several teams of researchers have prepared quasicrystalline samples perfect enough to give extremely sharp X-ray images that settle the question.

A normal crystal consists of groups of atoms that appear, like building blocks, in a regularly repeating pattern. In contrast, according to the original quasicrystal model, quasicrystals have at least two different basic building blocks, or unit cells, fitted together so as to create a structure that is neither regular nor random. Yet despite the lack of a perfectly repeating pattern, the orientation of one unit cell still determines the orientation of cells far away. The whole structure has a kind of long-range order (SN: 7/16/88, p.42).

But X-ray scattering experiments on the earliest known quasicrystalline materials produced images that were fuzzy, indicating the presence of more disorder than the quasicrystal model allows. Scientists favoring the quasicrystal model attributed the fuzziness to the presence of defects known as phasons, which correspond to misalignments of the material's unit cells.

However, Peter W. Stephens, presently at Tohoku University in Sendai, Japan, and Alan I. Goldman of Iowa State University in Ames suggested another possibility, which they called the icosahedral glass model. They proposed that the materials are more like glasses than defect-strewn crystals. In their model, groups of atoms form into a single type of unit cell in the shape of an icosahedron (a regular geometric figure having 20 triangular faces). Such units then fit together as best they can to create a rather sloppy but still partially ordered structure.

The discovery of a new class of quasicrystalline materials in Japan in 1987 set the stage for a test of the two competing models. Researchers found that combining aluminum and copper with either iron or ruthenium leads to crystals many times more perfect than any previously known quasicrystalline materials. Samples of these new materials, produced during the last two months at the IBM Thomas J. Watson Research Center in Yorktown Heights, N.Y., Tohoku University, AT&T Bell Laboratories in Murray Hill, N.J., and Harvard University, have none of the characteristic disorder seen in other quasicrystals.

"The materials are too perfect to be described sensibly in terms of an icosahedral glass model," says Paul J. Steinhardt of the University of Pennsylvania in Philadelphia. Steinhardt and Dov Levine proposed the original quasicrystal model.

"With these new materials, you can nail down the widths and positions of [X-ray] peaks to such a high accuracy that there's no need to discuss defects in the material," says IBM's Peter A. Bancel. "It really blows any competing picture out of the water."

"I'm tremendously excited," says Goldman, who is studying samples of the ruthenium alloy. "Of course, I'm a little disappointed that our model doesn't hold up here, but models are meant to be looked at critically and put aside if they don't explain the phenomena."

But many questions remain. Although researchers now know that true quasicrystals can be produced, no one knows why the new materials work so well and why the previously discovered materials form into such poor crystals. Researchers are now busy working on that puzzle.