Atom clusters act bigger than their britches.

Atom clusters act bigger than their britches

Half a dozen atoms do not a semiconductor make--in theory, at least. But two physical chemists now describe a "totally astonishing" result that leads them to question the current theory.

Tiny clusters of indium phosphide atoms -- which should behave much differently than the bulk material -- show optical properties resembling those of the semiconducting bulk form, report Kirk D. Kolenbrander and Mary L. Mandich of AT&T Bell Laboratories in Murray Hil, N.J. While scientists offer widely varying interpretations of the experiment, some suggest that further investigations of this phenomenon might have far-reaching implications for semiconductor technology.

Theoretically, for a material to behave like a bulk solid, it must contain many atoms together. A group of two or three iron atoms, for instance, won't conduct electricity like a large iron rod. To understand the transition between atomic and bulk behavior, scientists have been studying clusters ranging from hundreds of thousands of atoms to fewer than 10.

Mandich and Kolenbrander say their experiment is the first infrared spectroscopic investigation of six- to 12-atom clusters. Theoretical considerations and experiments with other materials led the team to expect large optical differences between the indium phosphide clusters and the bulk solid. Instead, the clusters' infrared spectra revealed photon absorption at approximately the same energy ranges as those of the bulk solid, they report in the Oct. 22 PHYSICAL Review LETTERS.

According to theory, electons have energies in a range, or continuum, of permitted values. In contrast, electrons in individual atoms or in simple, nonsolid molecules have discrete energy levels. In a semiconductor, the term continuum band refers to the particular band of electron energies at which the material conducts electricity. An electron in a semiconductor will not absorb photons with energies that would put it into a forbidden region, or band gap, directly below the continuum band, but it will absorb photons with any energy that wouldput it into the conducting region.

Kolenbrander (now at MIT) and Mandich say they aren't sure why their indium phosphide clusters showed a continuum absorption band rather than distinct absorption levels. They speculate, however, that the tiny clusters must have some electron bonds similar to those in the semiconducting solid.

If they're right, the finding may represent a step toward molecular-scale semiconducting devices. In recent years, scientists have been considering indium phosphide as one possible replacement for silicon in microscopic electronic circuits. "Theory says ... that if [clusters] get too small, they lose semiconducting properties. . . . But if the [band] is still there, it may be possible to make molecular-size integrated circuits, much smaller than anyone can do now," says chemist Michael A. Duncan of the University of Georgia in Athens.

However, the experimental result also has many possible explanations that do not involve semiconductivity, Duncan and others caution. There's no particular reason to believe that the band in the cluster has the same cause as the band in the solid, they say.

Mandich speculates that further studies of the cluster phenomenon may ultimately force physicists to revise their theories of amorphous solids -- a class that includes some forms of semiconductors -- to explain how such a tiny cluster can share properties with its bulk counterpart while apparently differing greatly in geometric structure.

She now plans to investigate other semiconductors, such as silicon, and slightly larger clusters of indium phosphide to see if they show similar spectra.