Phosphorus is a limited resource, and wastewater treatment plants (WWTPs) offer a potential site for its recovery, particularly as vivianite in Fe-based chemical phosphorus removal systems. However, vivianite formation along wastewater treatment trains is not yet fully understood. This thesis investigates this issue with a focus on sulphur (S) and phosphorus (P) competition over iron (Fe) employing full-scale WWTP data, tailored experiments for monitoring Fe redox states, and modelling tools. Full-scale tracking of vivianite was carried out at two large European WWTPs, Viikinmäki (Finland) and Seine Aval (France). Sampling of the water and sludge lines showed that Fe reduction and initial vivianite precipitation occur already in thickened and mixed sludges, within 1-2 days under anaerobic conditions. Full-scale Fe and P mass balances confirmed nearly complete Fe reduction prior to digestion and highlighted the importance of the Fe:P and Fe:S ratios, solids retention time and redox conditions on vivianite formation. The site sampling highlighted the limitation of the sole use of Mössbauer spectroscopy for vivianite quantification. This motivated the following study to develop a multi-method analytical framework combining chemical Fe2+/Fe3+ speciation, 57Fe Mössbauer spectroscopy at 300 K and 5.6 K, and X-ray diffraction. This approach enabled differentiation of vivianite Fe2+/Fe3+ from other Fe phases such as ferrihydrite, iron sulphides, and siderite which in turn improved potential phosphorus recovery in form of vivianite quantification with estimated shares of 20-71% in thickened sludge and 46-100% in digested sludge.
To obtain a comprehensive view of vivianite formation in WWTP, a model of Viikinmäki was set up and calibrated in SUMO© integrating Fe redox kinetics and Fe-P-S interactions along with the precipitation of vivianite, iron sulphides, and hydrous ferric oxides (HFO), and extended to include siderite. A first order Fe reduction rate of 0.80 ± 0.10 d-1, obtained from anaerobic batch tests experiments, was used to simulate HFO reduction. The calibrated model adequately estimated Fe2+/Fe3+ ratios and vivianite fractions. However, effluent total Fe was overestimated when using an oxidation rate obtained from literature warranting further work to understand Fe dynamics. To this end, ferrous iron dynamics were quantified experimentally under aerobic and anoxic activated sludge conditions. The results suggest that Fe2+ is not only consumed from liquid phase by oxidation under aerobic conditions, but also through other pathways, most likely rapid adsorption (<1 min) onto biomass. The anoxic Fe2+ oxidation rate was measured for the first time under typical activated sludge conditions and was only 2.7% of the aerobic oxidation rate at comparable total suspended solids (TSS). The resulting apparent aerobic rate constant (850 d-1 at 3.4 g TSS/l) was implemented in the full plant model, improving effluent Fe predictions without affecting effluent P, because phosphate concentrations were already low. However, this new rate increased the siderite precipitation in the sludge line.
Overall, the thesis sheds light on the overlooked Fe phases, demonstrates vivianite formation potential in the upstream of digesters and identifies key operational conditions for modelling, designing and optimizing phosphorus recovery strategies based on vivianite.