Modelling Cu(II) adsorption to ferrihydrite and ferrihydrite-bacteria composites: deviation from additive adsorption in the composite sorption system

Bacterially associated iron (hydr)oxides are widespread in natural environments and are potent scavengers of dissolved metal ions. However, it is unclear whether metal sorption on these composites adheres to the additivity principle, and thus whether metal concentrations in environments where these composites comprise a significant proportion of the reactive iron phases can be modelled assuming component additivity. Here we address this issue for Cu adsorption on ferrihydrite-Bacillus subtilis composites. We precipitated pure ferrihydrite and ferrihydrite composites with different ferrihydrite:bacteria mass ratios, and measured Cu adsorption as a function of pH, Cu adsorbed concentration and composite mass ratio. We develop a molecular-level surface complexation model for Cu adsorption on pure ferrihydrite. We then combine our end-member models for Cu adsorption on B. subtilis (Moon and Peacock, 2011) and ferrihydrite to model the observed Cu adsorption on the composites, adopting a component linear additivity approach. By comparing observed Cu adsorption to that predicted by our composite model, constrained to the exact best fitting end-member stability constants, we find that Cu adsorption behaviour on ferrihydrite-B. subtilis composites deviates from additivity. Specifically, Cu adsorption on composites composed mainly of ferrihydrite is enhanced across the adsorption pH edge (pH ∼3-6), while on our composite composed mainly of bacteria adsorption is enhanced at mid-high pH (pH ∼5-6) but diminished at mid-low pH (pH ∼5-3), compared to additivity. In current surface complexation modelling constructs, Cu adsorption on composites composed mainly of ferrihydrite can be modelled in a component additivity approach, by optimising the stability constants for Cu adsorption on the ferrihydrite and bacteria fractions to values that are within the uncertainty on the end-member stability constant values. The deviation from additivity of these composites, apparent when using the exact best fitting end-member stability constants, is therefore either a modelling artefact due to uncertainties in the surface adsorption properties of the end-member phases, or is relatively minor and cannot be separated from these uncertainties that are inherently present in the parameterisation of the surface adsorption properties in the modelling. In contrast, composites composed mainly of bacteria express significant deviation from Cu adsorption additivity that cannot be modelled in a component additivity approach. We propose that the deviations in Cu adsorption from additivity are the result of physiochemical interactions between the composite fractions that change the surface charge of the ferrihydrite and B. subtilis fractions compared to the isolated end-member phases. The magnitude and direction of the additivity deviations due to these electrostatic effects are then a function of the affinity of Cu for the bacterial fraction and the mass fraction of bacteria in the composite.