The adsorption of As(III) and As(V) on goethite has been studied as a function of pH and loading. The data can be successfully described with the charge distribution (CD) model (extended Stern layer option) using realistic species observed by EXAFS. The CD values have been derived theoretically. Therefore, the Brown bond valence approach has been applied to MO/DFT optimized geometries of a series of hydrated complexes of As(III) and As(V) with Fe(III) (hydr)oxide. The calculated ionic CD values have been corrected for the effect of dipole orientation of interfacial water, resulting in overall interfacial CD coefficients that can be used to describe the surface speciation as a function of pH and loading. For As(III), the main surface species is a bidentate complex and a minor contribution of a monodentate species is found, which is in agreement with EXAFS. The CD values have also been fitted. Such an analysis of the adsorption data resulted in the same surface species. The fitted CD values for the bidentate complex points to the presence of strong As-O bonds with the surface and a weaker As-OH bond with the free OH ligand. This agrees quantitatively with the MO/DFT optimized geometry. Interpretation of free fitted CD values for As(V) binding suggests that the main surface species is a non-protonated bidentate complex (B) with a contribution of a singly protonated surface complex (MH) at sub-neutral pH and high loading. In addition, a protonated bidentate surface complex (BH) may be present. The same species are found if the theoretical CID values are used in the data analysis. The pH dependency of surface speciation is strongly influenced by the charge attribution of adsorbed species to the electrostatic surface plane while the effect of loading is primarily controlled by the amount of charge attributed to the 1-plane, illustrating the different action of the CD value. The MO/DFT geometry optimizations furthermore suggest that for As(V) the B, MH and BH surface complexes may have very similar As-Fe distances which may complicate the interpretation of EXAFS data. (c) 2006 Elsevier Inc. All rights reserved.