Predicting nucleation and growth processes: Atomistic modeling of metal atoms on ionic substrates

Classical simulation methods are used to predict the energies required by nucleation models to explain the growth behavior of silver and gold on ionic substrates. Adsorption $E_a$ and diffusion $E_d$ energies are found to be small, with ${(E}_a-E_d)$ systematically smaller than several experimental estimates. Dimers have a large binding energy $E_b,$ but may have smaller diffusion energies than single adatoms. We show that surface point and line defects adsorb metal atoms, and discuss the implications of these results for analyses of nucleation data.