To prevent eutrophication of surface water, phosphate needs to be removed from sewage. Iron (Fe) dosing is commonly used to achieve this goal either as the main strategy or in support of biological removal. Vivianite (Fe(II) ₃ (PO ₄ ) ₂ * 8H ₂ O) plays a crucial role in capturing the phosphate, and if enough iron is present in the sludge after anaerobic digestion, 70–90% of total phosphorus (P) can be bound in vivianite. Based on its paramagnetism and inspired by technologies used in the mining industry, a magnetic separation procedure has been developed. Two digested sludges from sewage treatment plants using Chemical Phosphorus Removal were processed with a lab-scale Jones magnetic separator with an emphasis on the characterization of the recovered vivianite and the P-rich caustic solution. The recovered fractions were analyzed with various analytical techniques (e.g., ICP-OES, TG-DSC-MS, XRD and Mössbauer spectroscopy). The magnetic separation showed a concentration factor for phosphorus and iron of 2–3. The separated fractions consist of 52–62% of vivianite, 20% of organic matter, less than 10% of quartz and a small quantity of siderite. More than 80% of the P in the recovered vivianite mixture can be released and thus recovered via an alkaline treatment while the resulting iron oxide has the potential to be reused. Moreover, the trace elements in the P-rich caustic solution meet the future legislation for recovered phosphorus salts and are comparable to the usual content in Phosphate rock. The efficiency of the magnetic separation and the advantages of its implementation in WWTP are also discussed in this paper.

Arsenic is mainly removed from groundwater by adsorption onto, or co-precipitation with iron flocs and -deposits. The efficiency of this process depends on various factors, amongst which the oxidation state of arsenic, adsorption competition or poor iron floc removal. The aim of this study was to assess the adsorption efficiency of arsenic in full-scale groundwater treatment plants in the Netherlands. Adsorption efficiency is dependent on the concentration of adsorbent (Fe) and adsorbate (As), both of which vary considerably at the various treatment plants. To allow comparing treatment plants at these different Fe and As concentrations, the framework of a linearized isotherm graph was used. As a reference, jar tests were executed to derive adsorption isotherms of As(III) and As(V) based on co-removal with Fe²⁺ precipitation. All treatment plants have a higher adsorption efficiency than the As(III) reference, and lower adsorption efficiency than the As(V) reference. Treatment plants with an efficiency close to the As(III) reference may suffer from incomplete arsenic oxidation, although other causes cannot be ruled out. Classifying removal efficiency can be used in an initial research phase to identify treatment plants with relatively poor performance as candidates for improvement. Furthermore, treatment plants can be grouped into categories with similar adsorption efficiency, aiding in identification of common factors responsible for As removal.

The combination of ozonation and activated carbon (AC) adsorption is an established technology for removal of trace organic contaminants (TrOCs). In contrast to oxidation, reduction of TrOCs has recently gained attention as well, however less attention has gone to the combination of reduction with AC adsorption. In addition, no literature has compared the removal behavior of reduction vs. ozonation by-products by AC. In this study, the effect of pre-ozonation vs pre-catalytic reduction on the AC adsorption efficiency of five TrOCs and their by-products was compared. All compounds were susceptible to oxidation and reduction, however the catalytic reductive treatment proved to be a slower reaction than ozonation. New oxidation products were identified for dinoseb and new reduction products were identified for carbamazepine, bromoxynil and dinoseb. In terms of compatibility with AC adsorption, the influence of the oxidative and reductive pretreatments proved to be compound dependent. Oxidation products of bromoxynil and diatrizoic acid adsorbed better than their parent TrOCs, but oxidation products of atrazine, carbamazepine and dinoseb showed a decreased adsorption. The reductive pre-treatment showed an enhanced AC adsorption for dinoseb and a major enhancement for diatrizoic acid. For atrazine and bromoxynil, no clear influence on adsorption was noted, while for carbamazepine, the reductive pretreatment resulted in a decreased AC affinity. It may thus be concluded that when targeting mixtures of TrOCs, a trade-off will undoubtedly have to be made towards overall reactivity and removal of the different constituents, since no single treatment proves to be superior to the other.