Abstract
The 2×8 reconstruction of Ge films of Si(100) is shown to arise from ordered arrays of rebonded missing dimers (RMD’s). Such a reconstruction was suggested by Pandey for Si(100). However, because of their large tensile stress, RMD’s are much more energetically favorable in epitaxially compressed Ge films. For Si(100), RMD’s have a very small energy, and their presence may account for the puzzling discrepancy between theory and experiment for the stress anisotropy.