Collapsing clusters lead to fullerenes.

Chemists have quite successfully cooked up large quantities of fullerenes for three years now, but no one yet knows how these structures manage to emerge out of the hot carbon chaos. Why the commonly used arc-reactor-synthesis method works at all still mystifies researchers. How could atomized carbon spontaneously yield such highly ordered molecular cages?

New experimental evidence suggests that at high temperatures large carbon clusters form and then collapse into a more stable fullerene configuration.

"We've shown how carbon in a very high-energy environment reacts with itself and goes on to form fullerenes," says Michael T. Bowers of the University of California, Santa Barbara. "It's not what people -very reasonably-thought in the past."

The smallest observed fullerenes, containing 30 carbons, had appeared to come out of nowhere, Bowers says. He and his co-workers set out to discover how they form. Using a method called ion chromatography- which they developed to study carbon clusters - the team first determined what structures carbon atoms prefer to adopt.

Researchers had theorized that fullerenes assemble from sheets of pentagons and hexagons, but the group found no evidence of this. Instead, they observed that a few carbon atoms will link up linearly and that 10 carbons form monocyclic rings, 20 or more carbons form bicyclic rings, and 30 or more carbons form tricyclic rings.

In the May 6 NATURE, the California group describes how heating these large planar rings causes them to rearrange into the three-dimensional, spherical fullerenes. The rings melt down and a small carbon fragment evaporates as the atoms settle into their new arrangement.

Bowers speculates that in the searing carbon soup of an arc reactor, the 60carbon buckyball forms preferentially because it is the most stable fullerene in the intermediate size range. Larger fullerenes may coalesce just outside the arc's hottest region, where negatively charged carbon clusters may lose electrons and grow further before melting into fullerenes.

Robert E Curl of Rice University in Houston applauds the work for contributing to a fundamental understanding of fullerenes and for opening up theory to experimental testing. "Here's something that may bear very strongly on the formation of C60 and fullerenes in general:' he says.

Such research may one day help chemists control the kind of fullerene they produce. Says Bowers, "Once you know the mechanism, you have a chance at tailoring molecules. Until then, people are just playing in the dark."