Removal of Per- and Polyfluoroalkyl substances by anion exchange resins: Scale-up of rapid small-scale column test data.

Anion exchange (IX) is a readily implementable water treatment method that can effectively remove per- and polyfluoroalkyl substances (PFAS). The overarching objective of this research was to predict PFAS removal in full- or pilot-scale packed-bed IX resin contactors from rapid small-scale column test (RSSCT) data. Specific objectives were to (1) assess the effects of IX resin crushing on total anion exchange capacity and packed bed density, (2) determine the effects of initial PFAS concentration on PFAS uptake capacity, (3) determine the rate-limiting step controlling PFAS uptake kinetics, (4) determine the effects of hydraulic loading rate on PFAS uptake capacity, and (5) link constant diffusivity RSSCT data to pilot test data to develop a scale-up protocol. Experiments were conducted with two single-use IX resins and three water matrices, including coagulated surface water and groundwater. Crushing IX resin did not substantially change the bed density and total anion exchange capacity, but the morphology of particles changed from almost perfectly spherical to irregularly shaped. PFAS uptake capacity was independent of influent PFAS concentrations in the 30-300 ng/L range. This finding facilitated the development of an RSSCT scale-up approach because influent PFAS concentrations in RSSCTs and corresponding pilot tests often differ. Biot number values and data from interrupted RSSCTs demonstrated that film diffusion or a combination of film diffusion and intraparticle diffusion controls the rate of PFAS uptake by IX resins. From RSSCTs with identical empty bed contact times but different hydraulic loading rates (vf ), PFAS uptake capacity was found to be a function of the square root of the product of Sherwood number and particle shape factor (Sh×ϕ). Using a constant diffusivity RSSCT design with a reduced vf , full- or pilot-scale PFAS breakthrough data can therefore be predicted by multiplying the bed volumes of water treated in the RSSCT by a factor of (Shpilot ×ϕpilot )/(ShRSSCT ×ϕRSSCT ) . This research will support the design of future IX treatment processes in the context of PFAS remediation and drinking water treatment.