Marko Pranic

The removal of per- and polyfluoroalkyl substances (PFAS) by activated carbon (AC) and the regeneration of AC under conditions relevant for drinking water treatment were investigated in this PhD Thesis. The work was motivated by increasingly stringent regulatory limits for PFAS in drinking water, the growing relevance of short-chain PFAS, and the low removal of PFAS by granular activated carbon (GAC) filters operated at trace concentrations. By combining systematic batch experiments, mechanistic modelling, and regeneration studies, this Thesis provides a coherent framework to understand why PFAS adsorption is fundamentally limited at ng/L levels and how these limitations can be mitigated in practice.

A central finding of this Thesis is that PFAS adsorption on AC behaves fundamentally differently at environmentally relevant concentrations than at the elevated concentrations that dominate the scientific literature. Adsorption isotherm parameters were shown to depend strongly on the initial PFAS concentration used to reach equilibrium. Isotherms derived at µg/L–mg/L levels substantially overestimate adsorption capacity at ng/L concentrations, primarily because high concentrations promote cooperative adsorption phenomena such as aggregation and multilayer formation that are suppressed at trace levels. As a result, adsorption data obtained at high concentrations cannot be directly extrapolated to predict GAC performance in drinking water treatment.

PFAS adsorption kinetics were inherently slow under conditions relevant to drinking-water treatment. Intraparticle transport was shown to be governed predominantly by surface diffusion rather than pore diffusion, with surface diffusion coefficients decreasing when more adsorption sites are available, at lower PFAS concentrations, and in carbons with higher mesopore volume. As a consequence of surface diffusion controlling PFAS intraparticle transport, long equilibration times are required in batch systems and long empty bed contact times are required in column systems to achieve high PFAS removal.

Competitive adsorption between different PFAS was identified as a key mechanism controlling PFAS removal in GAC systems. Long-chain PFAS consistently outcompete short-chain PFAS due to stronger hydrophobic interactions, and in some cases, desorption of short-chain PFAS. This competition occurs both dynamically, due to differences in adsorption kinetics, and at equilibrium under conditions of limited adsorption space.

The role of AC properties was clarified by decoupling pore size distribution, surface chemistry, and transport effects. A higher micropore surface area did not result in increased PFAS adsorption. Instead, mesopores dominate PFAS adsorption because slow intraparticle transport prevents effective utilization of micropores. Carbons with high mesopore volume were identified as most effective for PFAS removal. Surface charge further influences adsorption: carbons with a low concentration of acidic surface groups or a net positive surface charge exhibit enhanced PFAS adsorption, particularly for short-chain compounds.

Natural organic matter (NOM) was identified as the dominant external factor limiting PFAS adsorption in water matrices often found in drinking water treatment. NOM reduced PFAS adsorption capacity by at least an order of magnitude, with low-molecular-weight NOM fractions identified as the most competitive species. These findings explain the rapid PFAS breakthrough observed in GAC filters and demonstrate that effective PFAS removal cannot be achieved by GAC alone without addressing NOM competition.

Temperature effects further limited PFAS removal. Increasing temperature reduced adsorption of short-chain PFAS, reflecting a lower affinity of these compounds for activated carbon at higher temperatures, whereas adsorption of long-chain PFAS was largely temperature-independent. As a result, short-chain PFAS removal performance may vary seasonally, which should be considered in treatment system design.

To translate these mechanistic insights into practice, this Thesis evaluated strategies to improve PFAS removal and reduce operational costs of drinking water treatment. The results indicate that selecting GACs with high mesopore volume and favourable surface charge, removing competing NOM through biological treatment or combined O₃–BAC processes, and applying predictive modelling frameworks that explicitly account for adsorption site competition and surface-diffusion-limited transport are the most promising near-term approaches.

Finally, this Thesis explored electroregeneration as a potential alternative to frequent thermal reactivation of GAC filters. Electroregeneration was shown to preferentially desorb short-chain PFAS and concentrate them into a smaller volume, demonstrating its potential to extend filter lifetime and reduce the environmental footprint of PFAS removal. While further development is required before full-scale implementation, these results provide a proof of concept for a more sustainable regeneration strategy.

In conclusion, this Thesis demonstrates that PFAS removal by activated carbon at drinking water concentrations operates at the limits of adsorption. Low influent concentrations, slow intraparticle transport, strong competition from NOM and co-occurring PFAS, and temperature effects collectively limit adsorption performance. By identifying these limitations and linking these to measurable material properties and operating conditions, this work provides a mechanistic basis for improving PFAS treatment design, modelling, and regeneration of activated carbon filters, while clearly defining the conditions under which activated carbon can, and cannot, adsorb PFAS effectively.