Ferlian Vida Satriaji, Ming-Chieh Li, Piotr Jachimowicz, H M Solayman, Ana Maria Janet Pacheco Mori, Cat Tuong Le Tong, Anggraini Widyastuti, Thanh-Binh Nguyen, Yoshifumi Horie, How-Ran Chao, Jheng-Jie Jiang
Nine target per- and polyfluoroalkyl substances (PFAS) were analyzed in aqueous samples from a full-scale municipal water resource recycling center using an anaerobic-anoxic-oxic process followed by a membrane bioreactor (MBR). Samples were collected from the influent, primary sedimentation, final sedimentation, MBR, and final effluent during 12 monthly grab-sampling events from June 2018 to May 2019 and a separate seven-day daily campaign from 21 to 27 May 2019. Seven compounds were detected above their compound-specific method detection limits in at least one sample, whereas PFUnA and PFDoA were not detected. In the monthly dataset, mean concentrations of the seven quantified compounds (ΣPFAS) were 20.84 ng/L in the influent, 29.62 ng/L after the MBR, and 26.21 ng/L in the final effluent. PFOA and PFOS were the largest contributors and together accounted for 70.4% of the estimated annual ΣPFAS discharge of 0.314 kg. Friedman tests showed significant differences in ΣPFAS among the five treatment stages in both the monthly and seven-day datasets (p = 0.002 and 0.013, respectively), although the direct influent-effluent comparison was not significant in either campaign. Compound-specific risk quotients ranged from 1.30 × 10-9 to 1.65 × 10-5 and were classified as negligible under the criteria used in this study. Because precursors, sludge, suspended solids, and total organic fluorine were not measured, the observed stage-wise differences cannot establish precursor transformation, phase partitioning, or a complete PFAS mass balance. The results provide facility-specific estimates of aqueous PFAS concentrations, treatment-stage variation, and final-effluent discharge.