A simple quantum chemical theory of dissociative adsorption

Potential energy curves for the adsorption of a hydrogen atom on the (100), (110), (111) and the stepped (311) crystal faces of copper have been calculated in the pairwise additive model for gas atom-solid interactions. A Morse function is used to represent the lowest singlet pairwise H-Cu interaction potential and its parameters are adjusted so that the calculated maximum bond energies conform with the available experimental data. Bond strength on the low index faces is found to increase with the adatom's local coordination number. Except for the edge sites, the steps on the (311) surface strengthen the bonds to sites on the component low index facets. A systematic study of the convergence of the bond energy as a function of the number of solid atoms is reported. The H-Cu(s) potential is shown to be relatively insensitive to changes in the first layer separation distance of the size inferred from experiment. A new model for diatom-solid potentials is proposed in which the diatom-solid potential is expressed as a sum of the differences between diatom-solid atom London-Eyring-Polanyi-Sato three-body potentials and the diatom singlet potential. This model is used to calculate potential curves for various approaches of a hydrogen molecule towards the same copper faces. H₂ is predicted to be physisorbed on each face. The atom-solid and diatom—solid potentials are used in conjunction with a model formulated by Lennard-Jones to estimate activation energies for dissociative adsorption. The correct order is obtained for the activation energies on the low index faces. Substantially lower activation energies are obtained for approaches toward many of the sites on the two low index facets of the (311) surface as compared to the same approaches towards the individual component faces. Dissociative adsorption is predicted to proceed without activation near the steps on this surface. In general, higher activation energies are obtained when the admolecule is perpendicular to the surface or facet in question. The simple idea that the activation energies are determined by small shifts of the atomic potential relative to the less structure-sensitive molecular potential works well for the low index faces, but is not wholly satisfactory for the stepped (311) surface. All results reported in this paper are negligibly different from those that are fully converged with respect to cluster size in the present model.