Biosolids are the solid byproduct resulting from the treatment of domestic sewage in a treatment facility. Biosolids contain large amounts of nutrients such as C and N making them an excellent fertilizer; however, they also contain trace amounts of heavy metals that can leach to the ground limiting their application rate. The leaching process of heavy metals from biosolids is dictated by the physical properties of the soil and by the solid/liquid partitioning of the metals. Biosolids contain multiple sorptive surfaces such as organic matter, iron, aluminum and manganese oxides, silicates and carbonates. To accurately predict leaching of metals from biosolids, the interaction of metals with these surfaces need to be considered. Beforehand, models have been developed to simulate the interactions of metals with individual sorptive surfaces such as hydrous ferric oxides and manganese oxides. The goal of this research was to develop a multisurface geochemical modeling approach to predict the release of As, Cd, Cr, Cu, Mo, Ni, Pb and Zn from biosolids and to determine the affinity of heavy metals for the different sorptive sites present in biosolids. First, pH dependent leaching and isotherm experiments were conducted on biosolids. A multisurface approach was implemented using the NICA-Donnan model to incorporate organic matter (OM) as a sorbent. The generalized two layer model was used to incorporate iron, aluminum and manganese oxides. Selective chemical extractions were conducted to determine the concentration of available surface sites. The multisurface geochemical model required a large number of laboratory measured input values that demanded extensive laboratory analysis and had an associated uncertainty for which there was little knowledge on its impact to the uncertainty of the output. A sampling based global sensitivity analysis was used to relate model output variability and uncertainty with the uncertainty of the input.
The leaching pattern of the heavy metals showed strong pH dependence, similar to other waste materials. Overall, the model accurately predicted the release of metals over the pH range and the isotherms. The percentage of active dissolved organic matter (DOM) necessary to successfully model the leaching of metals under acidic conditions was significantly lower than under basic conditions; nevertheless, in the solution phase Cd, Cr, Cu, Ni, Pb and Zn complexes with DOM were predominant for the entire pH range. Organic matter (OM) was the predominant sorptive site in the matrix, however simulations of a case scenario in which OM was completely removed showed that biosolids still retained a large sorption capacity. The sensitivity analysis showed that the dissolved concentration of metals was not sensitive to variations of the input concentrations of: SO4-2, Na+1, NO3-1, Cl-1, Mg+2, K+1, F-1 and H4SiO4. In other words, the dissolved metal concentrations were not affected by the presence of SO4-2, Na+1, NO3-1, Cl-1, Mg+2, K+1, F-1 and H4SiO4. The dissolved metal concentrations leached from biosolids were sensitive to total metal concentrations, total sorptive sites available, and DOC, PO4-3, Al+3, Mn+2, and Fe+3 concentrations. The model uncertainty and sensitivity to the different input values varied with pH. Additionally, each metal input was only relevant for its own output suggesting that in these circumstances there was no competition effect among metals. The uncertainty of the output varied between 5 to 8 orders of magnitude depending on the metal.