Bond-counting arguments, which were supported by ab initio calculations, predicted a lower barrier for the so-called leap-frogging diffusion of Pt ad-dimers on Pt(110)-1 x 2 than for adatom diffusion or ad-dimer dissociation. This conflicted with experimental data, and it was suggested that this indicated that contaminants had an effect. In particular, it was noted that a single CO molecule preferred to bind to a Pt ad-dimer on top of one of the dimer adatoms. Binding at the dimer bridge was less favourable, by 0.11eV. The ad-dimer dissociation energy, starting with the CO in the on-top geometry, was 0.073eV. This value was close to the 0.07eV which had been predicted on the basis of a statistical analysis of scanning tunnelling microscopic recordings of dissociation and re-association. It was noted, while considering whether this agreement was a coincidence, that a dimer with a single CO molecule firmly bound to one of the Pt atoms should not look symmetrical, while the dimer images in some previous studies did. It was proposed that only a small barrier (0.16eV) impeded the displacement of a CO molecule from the favoured on-top binding site to the dimer bridge. Therefore, at 300K, a CO molecule was expected to hop frequently from one end of a dimer to the other; or to be dragged by a scanning-probe tip. An alternative explanation of the observed symmetry was that dimers always bound two CO molecules. However, calculations indicated that two CO molecules on an ad-dimer repelled strongly enough to destabilize it.

Ordering of Self-Diffusion Barrier Energies on Pt(110)-(1 x 2). P.J.Feibelman: Physical Review B, 2000, 61[4], R2452-5