Epsilon Open Archive

Abstract

Per- and polyfluoroalkyl substances (PFASs) have been ubiquitously detected in drinking water whichposes a risk for human exposure. In this study, the treatment efficiency for the removal of 15 PFASs was examined in a full-scale drinking water treatment plant (DWTP) in the City of Uppsala, Sweden, over aperiod of two years (2015-2017). Removal of the five frequently detected PFASs was influenced by the total operation time of granular activated carbon (GAC)filters, GAC type and surface loading rate. The average removal efficiency of PFASs ranged from 92 to 100% for “young” GAC filters and decreased to 7.0-100% for “old” GAC filters (up to 357 operation days, 29 300 bed volumes (BV) treated). Flow-rates were adjusted in two full-scale GAC filters of different operational age to examine the removal of PFAS and organic matter depending on GAC operational age and operating flow. The decrease inflow-rate by10 L s(-1) from 39 to 29 L s(-1) led to an average increase of 14% and 6.5% in total PFAS removal efficiency for an “old”(264 operation days, 21 971 BV treated) and a “young” GAC filter (63 operation days, 5 725 BV treated), respectively. A cost-analysis for various operation scenarios illustrated the dominating effect of treatment goals and costs for GAC regeneration on overall GAC operation costs. The unit costs for GAC filters ranged from 0.08 to 0.10 E(-3) water treated and 0.020-0.025Vm E(-3) water treated for a treatment goal of 10 ng L(-1)and 85 ng L(-1), respectively, for Sigma(11)PFAS. Furthermore, it was concluded that prolonging the GAC service life by lowering the flow-rates after reaching the treatment goal could lead to a 26% cost-deduction. The results and methods presented in this study give drinking water providers valuable toolsfor the operation of a full-scale treatment train for the removal of PFAS in contaminated raw water.