To determine solid‐state structures, the electron localization function (ELF) must be interpreted somewhat differently than for molecules. The example of the diamond structures of C (right, top), Si, Ge, and α‐ and β‐Sn (right, bottom) show clearly that ELF depicts the electronic changes in regional space as the covalent bond gives way to metallic bonding.

Bonds and lone electron pairs can be made “visible” when the electron density distribution is used to calculate the electron localization function (ELF). This paper presents a computer‐graphics image of ELF in colors which represent the extent of the localization (at the right a black‐and‐white picture of N2).

No direct SiSi bond is present in 1 in spite of the short interatomic distance; however, the long CC distance does not prevent weak bond formation in 2. These results are predicted by an analysis of the compounds with the electron localization function. The explanation for the short SiSi bond in 1 lies in the different sizes of C and Si.

42 1

Our recent pseudopotential results for Cu2, Ag2, CuH, and AgH [J. Chem. Phys. 79, 5532 (1983)] are refined in two ways: (i) the density-functional approximation for valence correlation is replaced by a CI treatment, and (ii) the dipole polarization potential describing core-valence correlation is supplemented by a quadrupole correction term.