Showing papers by "David Vanderbilt published in 1987"

Absence of large compressive stress on Si(111).

Abstract: The surface stress tensor is calculated from first principles for several Si(111)1\ifmmode\times\else\texttimes\fi{}1, 2\ifmmode\times\else\texttimes\fi{}1, and 2\ifmmode\times\else\texttimes\fi{}2 surfaces. Contrary to widely held expectations, only a weak compressive stress is found for the unreconstructed 1\ifmmode\times\else\texttimes\fi{}1 surface. Reconstructions involving \ensuremath{\pi}-bonded chains or adatoms are predicted to be under substantial tensile stress. The results indicate that relief of surface stress is not the driving force for the 7\ifmmode\times\else\texttimes\fi{}7 reconstruction. The stress is shown to be useful in prediction of the effect of externally applied strain, which can drive changes in the surface reconstruction pattern.

Model for the energetics of Si and Ge (111) surfaces

Abstract: The reconstructions observed on the annealed (111) surfaces of Si and Ge can be explained on the basis of a simple model of the surface energetics The driving force for the Si 7\ifmmode\times\else\texttimes\fi{}7 reconstruction is the formation of dimer-row domain walls, which order in a triangular pattern for topological reasons Adatoms play an incidental role only Chemical or strain-induced variations in the parameters of the model can lead to transitions between energetically competitive 7\ifmmode\times\else\texttimes\fi{}7, 5\ifmmode\times\else\texttimes\fi{}5, $c2\ifmmode\times\else\texttimes\fi{}8$, and other structures

Bond relaxation in Hg1−xCdxTe and related alloys

Abstract: The local structural properties of Hg1−xCdxTe, Hg1−xZnxTe, and Cd1−xZnxTe are examined using a self‐consistent pseudopotential approach. An accurate description of the limiting crystals is obtained by adding empirical corrections to ab initio total energies calculated with the group‐II d states included in the cores. Similar calculations for x=0.5 ordered alloys confirm the presence of nearest‐neighbor bond‐length differences in these systems without violation of Vegard’s law. An anomalously large ‘‘relaxation’’ is predicted here for Hg1−xCdxTe (∼2% bond‐length difference in the alloy compared to only 0.3% between limiting crystals) due to the dominance of chemical effects neglected from simple valence force models. The implications of the present results for alloy mixing energies and possible deviations from randomness are discussed.