Effects of PFAS-Antiscalant Interactions in Foam Fractionation

Abstract

PFAS are a growing concern for drinking water treatment because of their persistence, mobility and potential health risks. Reverse osmosis (RO) is an effective technology for PFAS removal, but it produces a concentrate stream in which PFAS and other dissolved compounds accumulate. This concentrate requires further treatment to prevent the PFAS problem from being transferred to another waste stream. Foam fractionation is a promising option for this type of stream because it uses the surface-active properties of PFAS. During foam fractionation, PFAS adsorb to the air-water interface of rising bubbles and are transported to the foam phase, producing a relatively small foamate stream with elevated PFAS concentrations.

Previous research reported high PFAS removal efficiencies during foam fractionation of membrane concentrate containing antiscalant. It was suggested that the antiscalant present in the concentrate may have contributed to this improved performance. Antiscalants are added to RO systems to prevent scaling and are largely retained in the concentrate stream. Because many antiscalants are polymeric or contain anionic functional groups, they may interact with PFAS, dissolved ions or the air-water interface. However, it remains unclear whether such interactions influence PFAS removal and enrichment during foam fractionation. Therefore, this study investigated the effect of different antiscalants on PFAS removal from RO concentrate.

All experiments were performed with RO concentrate from Dunea's UF/RO pilot. The concentrate was collected on one day and in one batch to minimise differences in water composition between experiments. Six commercial antiscalants were tested: 4Aqua OSM BD30, Ameroyal 363, Vitec 1141, Vitec 1200, Sokalan RO400 and Aquatreat 535. A control experiment without antiscalant was also performed. The antiscalants were dosed at 23.5 mg/L, based on the expected concentration in the RO concentrate when the RO system operates at 80% recovery. The foam fractionation experiments were carried out in a continuous column setup with a contact time of approximately 30 minutes and an airflow rate of 4 L/min. Samples were collected from the influent, effluent and foamate after steady-state operation had been reached. PFAS concentrations were measured using LC-MS. In addition, pH, conductivity, surface tension, viscosity and visual foam characteristics were evaluated.

The effect of antiscalants on foam formation and foam characteristics appeared limited. The surface tension of the control sample was 69.26 mN/m, while the antiscalant-containing samples showed slightly lower values between 66.73 and 67.96 mN/m. This indicates that antiscalants slightly lowered the surface tension of the RO concentrate. However, the differences between antiscalants were small. Viscosity values were almost identical for all samples, and visual observations did not show clear differences in foam height, bubble size or foam structure. Therefore, the tested antiscalants did not appear to strongly influence the visible foam characteristics under the applied conditions.

The mean sum PFAS removal efficiencies ranged from 64.8% to 79.2%. The control experiment without antiscalant showed a removal efficiency of 71.8%. Most antiscalant-containing experiments showed slightly higher removal efficiencies than the control, except for V1141, which showed the lowest removal efficiency. However, the differences between experiments were relatively small. The results therefore do not show a clear or dominant improvement in total PFAS removal due to antiscalant addition under the applied conditions.

Mean sum PFAS enrichment factors showed larger differences between experiments than removal efficiencies. The mean sum PFAS enrichment factors ranged from 2.48 to 6.86, with the highest values observed for BD30 and V1141. This suggests that antiscalants may have influenced the partitioning of PFAS between the bulk liquid and the foam phase. However, the experiments with the highest enrichment factors also showed larger deviations in mass balance closure. As a result, enrichment factors should be interpreted together with removal efficiencies and mass balance recoveries rather than as stand-alone performance indicators.

The compound-specific results showed that PFAS properties had a strong influence on foam fractionation performance. Long-chain PFAS were generally removed more efficiently than short-chain PFAS. This is expected because long-chain PFAS are more hydrophobic and more surface-active, causing them to adsorb more strongly to the air-water interface. Short-chain PFAS, such as PFBS, were more difficult to remove and showed lower enrichment. These results indicate that chain length and molecular structure are important factors in foam fractionation performance and may have a stronger influence than the specific antiscalant type.

The mechanism by which antiscalants may influence PFAS foam fractionation remains uncertain. Based on literature, possible mechanisms include changes in cation availability, electrostatic interactions, PFAS hydrophobicity, interfacial adsorption behaviour and foam film stability. Since most tested antiscalants are expected to be anionic, direct electrostatic attraction with anionic PFAS is not expected. Their effect is therefore more likely to be indirect and related to matrix effects. Such effects may be more relevant for short-chain PFAS, because their partitioning to the air-water interface is weaker and more sensitive to changes in water composition.

Several limitations should be considered when interpreting the results. Although one RO concentrate batch was used for all experiments, some matrix variability within this batch may still have occurred. In addition, PFAS analysis was performed in two batches. This may have contributed to analytical variability and is especially relevant for the interpretation of the V1141 experiment, for which the influent sample was analysed in a different batch from the effluent and foamate samples. The TOC data were also not sufficiently reliable to support a detailed interpretation of organic matter effects.

Overall, this study showed that foam fractionation can remove PFAS from RO concentrate, with removal efficiencies comparable to those reported in literature. The results suggest that antiscalants may have contributed to differences in PFAS enrichment and compound-specific behaviour, but no clear or dominant antiscalant effect could be confirmed. Antiscalants should therefore be considered as one possible matrix component that may influence foam fractionation performance, together with PFAS chain length, water composition, conductivity, foam stability and analytical uncertainty. Future research should include improved water quality characterisation, higher antiscalant concentrations and positive control experiments with a known cationic co-surfactant to better understand the mechanisms behind possible antiscalant effects.