Effects of recirculation in a three-tank pilot-scale system for pharmaceutical removal with powdered activated carbon

The removal of pharmaceutically active compounds by powdered activated carbon (PAC) in municipal wastewater is a promising solution to the problem of polluted recipient waters. Today, an efficient design strategy is however lacking with regard to high-level overall, and specific, substance removal in the large scale. The performance of PAC-based removal of pharmaceuticals was studied in pilot-scale with respect to the critical parameters; contact time and PAC dose using one PAC product selected by screening in bench-scale. The goal was a minimum of 95% removal of the pharmaceuticals present in the evaluated municipal wastewater. A set of 21 pharmaceuticals was selected from an initial 100 due to their high occurrence in the effluent water of two selected wastewater treatment plants (WWTPs) in Sweden, whereof candidates discussed for future EU regulation directives were included. By using recirculation of PAC over a treatment system using three sequential contact tanks, a combination of the benefits of powdered and granular carbon performance was achieved. The treatment system was designed so that recirculation could be introduced to any of the three tanks to investigate the effect of recirculation on the adsorption performance. This was compared to use of the setup, but without recirculation. A higher degree of pharmaceutical removal was achieved in all recirculation setups, both overall and with respect to specific substances, as compared to without recirculation. Recirculation was tested with nominal contact times of 30, 60 and 120 min and the goal of 95% removal could be achieved already at the shortest contact times at a PAC dose of 10-15 mg/L. In particular, the overall removal could be increased even to 97% and 99%, at 60 and 120 min, respectively, when the recirculation point was the first tank. Recirculation of PAC to either the first or the second contact tank proved to be comparable, while a slightly lower performance was observed with recirculation to the third tank. With regards to individual substances, clarithromycin and diclofenac were ubiquitously removed according to the set goal and in contrast, a few substances (fluconazole, irbesartan, memantine and venlafaxine) required specific settings to reach an acceptable removal.