Low-energy-electron dift'raction (LEED) I-V curves have been measured for H adsorbed onPd(111) in the (&3X&3)R30'-2H structure, which occurs at a coverage e= 3. The measurements are compared to calculations using dynamical LEED techniques. Twenty-one distinct geometry types were investigated, each one with metal-interlayer relaxations allowed and the distance of the H atom above (or below) the top Pd layer varied. The metal-interlayer spacings that we find are within error bars of those of the clean surface. Agreement between theory and experiment was achieved for structures with one of the two H atoms in the unit mesh in a threefold hollow surface site, above third-layer metal atoms (denoted "A +"). The second H resides either in the same type of site or in another type of threefold site, such as the hollow surface site above second-layer metal atoms B+ or the subsurface sites between firstand second-layer metal atoms, Aor B -. We find that good reliability factors R are achieved for subsurface occupation fractions up to 60%. The results are consistent with the embedded-atom-method prediction of substantial subsurface H and with previous electron-stimulated desorption measurements.

No abstract

No abstract

We present calculations, based on the embedded-atom method of phonon energies and polarizations on the clean Cu(lOO) surface. The atomic relaxations and the dynamical matrix are obtained, without fitting to surface properties, in one consistent calculation which is no more computationally demanding then pair potentials. Excellent agreement with experiment is obtained. We find that the force constants are qualitatively different from those obtained by fitting to two-body central-potential models, reflecting the many-body nature of the method and suggesting that fits based on central-potential models may be misleading.

No abstract