On the structure of the Si(103) surface

Abstract

Although (103) is a stable nominal orientation for both silicon and germanium, experimental observations revealed that in the case of silicon this surface remains disordered on an atomic scale even after careful annealing. We report here a set of low-energy reconstruction models corresponding to $1\times 2$, $2\times 2$, and $1\times 4$ periodicities, and propose that the observed disorder stems from the presence of several coexisting reconstructions with different morphologies and nearly equal surface energies. These models also suggest that the model structures previously reported in the literature for the (103) orientation have very high surface energies and are thus unlikely to be experimentally observed.