It was recalled that scanning tunnelling microscopy showed that the Si(100) surface had a large number of missing-dimer vacancies and almost no missing-atom vacancies. Here, a kinetic explanation was offered for these observations. The Stillinger-Weber potential was used to calculate the energy barriers which opposed various atomic and dimer displacements along the surface. These calculations indicated that vacancy formation via thermal fluctuations was a very slow process. If a single-atom vacancy was formed, the atom that was dimerized with the missing atom would rapidly leave the dimer row. The missing-atom vacancy was thus transformed into a missing-dimer vacancy. The probability that 2 adsorbed atoms would fill a missing-dimer vacancy was very low. That is, the first atom which filled the vacancy would leave it quickly, and its lifetime at the vacancy site was shorter than the time within which the second atom could join it. Also, vacancy-filling competed, for adsorbed single atoms, against capture by steps and islands and dimer formation on top of the surface.

Z.Zhang, H.Metiu: Physical Review B, 1993, 48[11], 8166-71