Vibrational properties of adsorbates and reconstructed surfaces

Abstract

Three problems were investigated: (1) Reconstruction of fee (100) surface. A few years ago, Burton and Jura showed that an Einstein model predicts a structural phase transition for the (100) surface of Ar at a temperature T (subscript c) below the melting point. One would expect this transition to be experimentally observable in other fcc materials--A1, Ni, etc. In this work, however, we prove that the Burton and Jura model for the reconstructed surface is mechanically unstable and the phase transition is consequently impossible. (2) Vibrational modes of adsorbed atoms. Previous studies of the vibrational spectra of adsorbate-substrate systems have involved "mass defect" models, in which the adsorbate is assumed to differ from the substrate only in mass. Our calculations are for the systems ⁴He on Xe; Ne on Ar, Kr, and Xe; Ar and Kr on Xe; and Kr on Xe; with the proper force constant changes at the surface taken into account. A detailed discussion is given of the vibrational spectra of these systems. We find that the results for "mass defect" models are qualitatively incorrect in two ways: the ordering of the "principal" modes associated with adsorbate vibrations is backwards (with the vertically polarized modes lying below the horizontally polarized modes), and the effect of the adsorbate on the substrate surface modes is too large in the case of weakly bound adsorbates. (3) Heat capacities of adsorption. For the seven systems listed above, we calculated the heat capacities of adsorption. The results for Ne on Ar and Kr are in semiquantitative agreement with the experimental data of Antoniou for Ne on graphite. The results are also in qualitative agreement with the experimental data of Bretz et a l. for He on graphite.