Filling transitions on rough surfaces: inadequacy of Gaussian surface models
We present numerical studies of wetting on various topographic substrates, including random topographies. We find good agreement with recent predictions based on an analytical interface-displacement-type theory \cite{Herminghaus2012, Herminghaus2012a}. The phase diagrams are qualitatively as predicted, but differently in this study the critical points are found to lie within the physical parameter range (i.e., at positive contact angle) in all cases studied. Notably, it is corroborated that Gaussian random surfaces behave qualitatively different from all non-Gaussian topographies investigated, exhibiting a qualitatively different phase diagram. This shows that Gaussian random surfaces must be used with great care in the context of wetting phenomena.