It had long been assumed that alkali metal adatoms (Na, K, Rb, Cs) on close-packed metal surfaces would occupy high-symmetry, highly coordinated sites on an almost unperturbed substrate and be arranged homogeneously on the surface with maximum adatom—adatom distance in order to minimize the repulsive dipole-dipole  interaction, induced by partial charge transfer from the alkali metal adatom to the substrate. Studies performed over the last 15 years,  however, have shown that this view is far from correct and top sites may be occupied, as well as unexpected substitutional sites, formation of surface alloys, and island formation, resulting in an updated new view of alkali metal adsorption. (See C. Stampfl and M. Scheffler, Surf. Rev. Lett. 2, 317  (1995) for a review.)

Example, (2x2)-2Na/Al(111)
Experimental studies had difficulty in determining the atomic structure of this system (below). Our density-functional theory calculations however, correctly predicted it. [J. Burchardt, M. M. Nielsen, D. L. Adams, E. Lundgren, J. N. Andersen, C. Stampfl, M. Scheffler, A. Schmalz, and J. Haase et al. Phys. Rev. Lett 74, 1617 (1995).]