Heavy metal adsorption on iron oxide and iron oxide-coated silica ...

26and Sigg, 1992; Gunneriusson et al. 1994; Nowack et al. 1996). Assumed surfacecomplexes are checked by the goodness of the model fit (Goldberg, 1991). For example,shifts in the <strong>adsorption</strong> edge at very low coverage are usually attributed to the presence ofa small number of high affinity sites (Honeyman, 1984). Dzombak and Morel (1990)used two types of sites, a high affinity and a low affinity one, to describe the pHdependence of <strong>adsorption</strong> data. Although these approaches are usually successful, thesurface species selected may not be unique or physically meaningful.Fortunately, recent advances in addressing speciation using in situ spectroscopictechniques such as XAS and FTIR spectroscopy have made it possible to overcome thisdrawback (Hayes and Katz, 1996; Brown, 1990). Especially from XAS analysis, innersphereversus outer-sphere complexation, mononuclear versus multinuclear, or solidsolution formation can be distinguished through a local structure analysis (e.g.,Chisholm-Brause et al. 1990; Charlet and Manceau, 1992; O'Day et al. 1994; Bargar etal. 1997a, b; Scheidegger et al. 1998; Manning et al. 1998; Sahai et al. 2000; Trainor etal. 2000; Nachtegaal and Sparks, 2002; Trivedi et al. 2003, 2004; Wazne et al. 2003).For example, vanadium (V) <strong>adsorption</strong> on goethite at pH 1.5 to 12 could be bettermodeled with two monodentate surface complexes as compared to two bidentate species;however, based on XAS, only inner-sphere bidentate surface complexes formed (Peacockand Sherman, 2004). Also, SCM modeling was found to under-predict Co(II) <strong>adsorption</strong>on a-Al2O3 at high surface coverage (> 10%) using mononuclear complexes; while basedon XAS results, Katz and Hayes (1995b) successfully modeled this <strong>adsorption</strong> usingpolymer complexes.