Surface-stress-induced order in SiGe alloy films

Ordering in epitaxial Si/Ge alloys grown by molecular-beam epitaxy has been observed by several authors. Attempts to explain this unusual phenomenon on the basis of the bulk thermodynamic properties of the alloy have failed. In this Letter we show that the observed ordering is not an equilibrium bulk property, but rather the result of surface growth kinetics. We attribute the ordering to atomic-scale stresses present in the reconstructed Si(001) surface during growth, leading to double-layer segregation of Ge and Si along one of the four equivalent 〈111〉 directions. This near-surface ordering is then quenched-in as the crystal grows.