Adsorbate-induced surface stress, surface strain and surface reconstruction: CH3S on Cu(100) and Cu(111)

Density functional theory (DFT) calculations have been applied to study the structural phases formed by CH3S on Cu(100) and Cu(111). On Cu(100), the results show that while the observed 0.25 ML (2 × 2) phase is stable, a faulted missing-row 0.33 ML c(6 × 2) is energetically preferred to a 0.5 ML c(2 × 2) phase, consistent with experimental observations. The calculations also show that a c(2 × 2) phase would have a large adsorbate-induced compressive surface stress that is not found in the c(6 × 2) phase. An alternative model of the c(6 × 2) phase as a buckled c(2 × 2) structure is shown to be unstable. On Cu(111), a 0.33 ML (√3 × √3)R30° structure on an unreconstructed surface is found to be less stable than a 43−13 reconstructed surface with the same coverage, consistent with experimental data that the (√3 × √3)R30° phase is only seen at low temperatures and transforms to the reconstructed phase on heating. DFT calculations show that the originally proposed pseudo-(100) structure for this reconstruction is heavily distorted, consistent with the results of other recent calculations. A range of alternative reconstruction models are found with closely similar total energies, the structural model with (marginally) the lowest energy having 20% less Cu adatoms in the reconstructed layer than in the ideal pseudo-(100) model. Simulations of medium energy ion scattering (MEIS) data for these different models show this lowest-energy model to give the best fit to the MEIS data, but the scattered-ion yield enhancement is sensitive to which of two alternative versions of this model, involving Cu adatoms predominantly in fcc or hcp sites, is occupied. The possible role of local disorder and structural variability in the surface, and whether the reconstruction could be incommensurate, is discussed.