It had been pointed out that there was a sub-surface self-interstitial atom, in each 3 x 2 unit cell of the (113) surface, analogous to the case of (113) Si surfaces. It was observed here, by means of scanning tunnelling microscopy, that such interstitial atoms migrated frequently on clean and well-annealed (113) Ge surfaces; even at room temperature. The energy barrier to migration was deduced to be 0.93eV by assuming that the migration was a simple thermally activated process with an attempt frequency of 1013Hz. Inward and outward migrations involved approximately equal energy barriers; thus implying that the 3 x 1 and 3 x 2 reconstructions had approximately equal energies. It was thought to be very likely that, during the entire process of inward or outward migration, the sub-surface interstitial atom never became a surface atom. It was noted that the scanning tunnelling microscope tip could cause an interstitial atom to migrate away when the tip was directly above the atom, and could cause an atom to migrate to a vacant sub-surface interstitial site when the tip was directly above the site.

Z.Gai, R.G.Zhao, W.S.Yang: Physical Review B, 1997, 56[19], 12303-7