Octahedra and Trigonal Prisms
In fact, the few hundred carbides known are a most remarkable and, as I will argue, inspiring group of compounds. First a brief on their macroscopic properties: These are all solid compounds, ceramic, sometimes metallic, often hard, often with high melting temperatures. They may be sensitive to moisture (Can anyone who has experienced it forget the smell of wet technical grade CaC2?), but often are very resistant to water and air (those WC drilling bits). Their compositions may be simple—say, WC or NbC (niobium carbide)—or complex—Sc3C4 (scandium carbide) or Er10Ru10C19 (an erbium ruthenium carbide).
Now to their microscopic structure. In the carbides known, carbon appears so far in only three forms. Most common are isolated carbon atoms, probably closer to being negatively charged ions. Figure 2, for instance, shows the structure of NbC—a layering of niobium and carbon, with an octahedron of niobiums around each carbon and an octahedron of carbons around each niobium. Does that sound familiar? These words describe the geometrical structure of rock salt or NaCl (sodium chloride), seen in our view of NbC from an atypical angle. The distances between niobium atoms are a little longer than in niobium metal; the Nb-C distances are similar to those in discrete molecular NbC organometallic compounds. And there are no C-C bonds. Figure 2 also shows the WC structure, in which each carbon is surrounded by a trigonal prism of tungsten atoms and each tungsten by a trigonal prism of carbons. Interesting that the two structures should differ so much; I've written a paper with Sunil Wijeyesekera teasing out an explanation for the different coordination preferences of NbC and WC. Cementite, Fe3C (an important form of carbon in steel), is a variant of the WC structure. To get to it, imagine removing two-thirds of the carbon atoms and then collapsing the metal lattice, albeit in a complex way.
Incidentally, if you ask an inorganic chemist about the date of the first synthesis of molecules with a metal atom in a trigonal prismatic environment, the likely answer will be approximately 1965. But the WC structure has been known since 1928; it clearly has a trigonal prismatic coordinated metal atom, with W-C distances typical of those in organometallic molecules. That the molecular inorganic chemist doesn't point to WC (or to MoS2—molybdenum sulfide—another structure with trigonal prismatic coordination of a metal, given to us by Linus Pauling in one of his early papers in 1923) illustrates the separation, persistent even as it is illogical, of molecular and extended inorganic chemistry over most of the 20th century. Talk of two cultures!