Neglected adsorbate interactions behind diffusion prefactor anomalies on metals

Abstract

Highly anomalous values for the preexponential factor in atomic diffusion rates at surfaces have recently been inferred from scanning-tunneling microscopy growth experiments. In an extensive first-principles kinetic Monte Carlo study, we show how long-range adsorbate interactions invalidate the standard nucleation-theory approach to analyzing experimental island-density data. When adatom-adatom interactions are properly accounted for in the analysis of experimental data, the anomaly is lifted, and deduced prefactors are consistent with direct theoretical calculations. We show that the dependence of the island density on the growth rate is modified by interactions, which could be used to identify adsorbate interactions in growth experiments.