The validity of the specific reaction rate constants (k(SA)) in modelling contaminant removal in Fe-0/H2O systems is questioned. It is shown that the current k(SA)-model does not consider the large reactive surface area provided by the in-situ formed oxide film, and thus the adsorptive interactions between contaminants and film materials. Furthermore, neither the dynamic nature of film formation nor the fact that the Fe-0 surface is shielded by the film is considered. Suggestions are made how the k(SA)-model could be further developed to meet its original goal. (C) 2008 Elsevier B.V. All rights reserved.