metals) are still problems for such uses. One approach to ceramics with better properties is to overcome their fragility by incorporating them into composites. As chemistry moves from pure materials to organized systems of different materials, composites are leading the way. A challenge for the future is to invent improved structural materials, probably composites based on resins or on ceramics, that are stable at high temperatures and easily machined.

Carbon atoms in pure form can be obtained as materials having two classic types of molecular structure: diamond and graphite. In diamond each carbon atom is linked by equivalent single bonds to four neighboring carbons. The result is a clear very hard material that is used for cutting, in saws with tiny diamonds imbedded in the blades, as tough coatings for metals, and in other industrial uses as well as in jewelry. By contrast, each carbon atom in graphite is linked to only three neighboring carbons, in a sheet, and some of the electrons are in delocalized pi orbitals that permit them to move easily along the sheets. The extensive aromaticity of carbon sheets leads to electronic transitions with energies in the visible light region, so that graphite absorbs throughout the visible region and is a black material. In addition, the mobility of the pi electrons in graphite makes it an electrical conductor, in stark contrast to the insulating properties of diamond.

A new type of structure has recently been discovered in which the sheets of graphite-like carbons are curved. The first example, called fullerene (after the geodesic domes of Buckminster Fuller), has 60 carbons in a sphere. It resembles a soccer ball with its five- and six-sided polygons (in contrast to graphite, which resembles a floor tile pattern, with hexagons only). A Nobel prize was awarded in 1996 to Robert F. Curl, Jr., Harold W. Kroto, and Richard E. Smalley for their discovery of fullerenes. Instead of curling into a sphere, the sheets of carbon with hexagons can also curve into tubes with diameters on the order of 1 nm (often called nanotubes), tiny whiskers that are sometimes quite long. Because of the electrical conductivity of pi electrons, these tubes are also electrically conducting, somewhat like graphite. While they are already used in research instruments to probe microscopic structures, one of the challenges is to use these new structures in miniature devices, or as building blocks for organized chemical structures.

The importance of crystal form often is underappreciated. In many applications—from drugs (in which bioavailability may be determined by crystal form) to explosives (where crystals may differ in stability) and optical devices (where the nonlinear optical properties required for the device are based on a particular crystalline architecture)—the correct crystalline form is essential to obtaining the desired chemical and physical properties of a material. Crystallization has long been an art rather than a science; sometimes the same substances will exhibit polymorphism and adopt different crystalline forms depending on the crystallization conditions. Crystal engineering—the prediction and control of molecular crystal structures based on the constituent molecular structures—is on the verge of becoming a science. The current generation of computers is finally powerful