Combined removal of phosphorus and organic micro pollutants by adsorption onto powdered activated carbon and coagulation

Abstract

Jar test experiments were collected to study the removal of phosphorus and of the dissolved chemical oxygen demand (CODf) at different coagulant dosages. Effluents from wastewater treatment plants (WWTPs) Waßmannsdorf, Münchehofe and Ruhleben, and influent water from the surface water treatment plant (SWTP) Tegel were tested for this study. The CODf removal depends on the coagulant dose. CODf was removed from 10%, applying a coagu-lant dose of 4 mg Fe/l, to 23% applying a 12 mg Fe/l dose. Removal of soluble phosphorus (sP) and soluble reactive phosphorus (srP) at different coagulant dosages was studied. The proportion of non-reactive phosphorus (snrP) in the soluble phosphorus fraction in the tested waters was regarded as a parameter that describes the wide range of removal rates found in jar test experiments. With the results collected from jar tests done with water from different WWTP’s effluents, an empirical model for the sP and srP removal was created. The sP and srP concentration, the coagulant dose and the snrP ratio were used as input parameters to simulate the sP and srP removal performance of a chemical phosphorus precipitation. Phos-phorus removal performance of SWTP Tegel was predicted using the empirical model. Co-agulation/flocculation and direct filtration process was also predicted for the effluent of WWTP Waßmannsdorf. Two simulations were done for this case, one using the empirical model for the phosphorus removal, and the other one using results from jar test experiments done with the WWTP effluent. Results from both simulations showed that applying a coagu-lant dose of 6 mg Fe/l or higher, total phosphorus concentrations below 0.1 mg/l at the efflu-ent of WWTP Waßmannsdorf can be reached. A pilot plant is put into operation in the SWTP Tegel to investigate with the combined remov-al of phosphorus and organic micropollutants (OMPs) by chemical precipitation and adsorp-tion onto powdered activated carbon (PAC). The pilot plant consists of a mixing cascade, a settling cyclone with recirculation of the sludge to the mixing cascade, and subsequent filtra-tion in a dual-media rapid filter. The removal performance of organic bulk parameters (COD, DOC, UV254, UV436), phosphorus and OMPs were analyzed among three different experi-mental phases in the pilot plant. The first phase was done adding coagulant doses of 10 and 12 mg Fe/l and polymer doses of 1 and 0.5 mg/l. Total phosphorus (TP) effluent concentra-tions of the pilot plant were below 0.05 mg/l operating at similar coagulant and polymer dos-ages as the SWTP, with an average removal rate of 92%. COD removal rate was on average 27% in these conditions. In the second experimental phase in the pilot plant, the mixing cas-cade and the settling cyclone were bypassed, performing a coagulation/flocculation and di-rect filtration process. Results for phosphorus removal in this phase were compared with re-sults from the pilot plant in WWTP Münchehofe, which uses the same process. TP removal applying same β (relation, in moles, of the added coagulant dose to the srP concentration at the influent), had different results for the pilot plants. Using β=4.4, TP removal in pilot plant was 68% and in pilot plant Münchehofe was 90%. Using the same coagulant dose (5 mg Fe/l) as in pilot plant Münchehofe (β=4.4), TP removal in Tegel was 92% (β=14.6), very comparable to the 90% of in Münchehofe. The third experimental phase was done with mix-ing cascade, sedimentation and filtration. 5 mg Fe/l and 20 mg PAC/l were dosed during this phase. Since the beginning of the phase 3, three samples were taken for OMP analyses with different times of operation. The results with approximately 1.5 g PAC/l in the system and 55 minutes of reaction time, were compared with jar test results done with 20 mg PAC/l and 2 h of reaction time. Removal rates were in the same range for all the analyzed substances. Re-sidual concentrations of 4 and 6% were found for good adsorbable compounds as the anti-convulsant carbamazepine and the analgesic diclofenac, but for the anticonvulsant gabapen-tine or the contrast agent diatrizoic acid there was almost no removal.