Sonolysis of per- and polyfluoroalkyl substances (PFAS) has recently matured to field studies, treating real world contamination. However, efficient sonolysis reactor designs are poorly researched. Moreover, the variety and complexity of PFAS pollution slows reactor optimisation and scale-up. In this work, the defluorination of 10.0 mg/L aqueous perfluorooctane sulfonic acid (PFOS) was used as a model metric for the optimisation of; reactor volume (0.6 or 1.4 L), power density (100 – 350 W L-1), number of modular reactors (1–3), and liquid height (56.7 – 340 mm). Note, the ultrasonic frequency (410 kHz) and flow rate (214.2 ml min−1) was optimised in this reactor previously. Peak PFOS defluorination rate (3.40 μmolL-1 min−1) occurred at 141.8 mm, in a 0.6 L reactor, under 200 WL–1 ultrasound. Increasing the number of transducers connected in parallel to one amplifier was able to increase treatment efficiency from 78.6 to 191.8 μmol kWh−1. The model was validated using legacy aqueous film forming foam (AFFF, 3 M FC-602 Lightwater) at different dilutions (×5, ×10, ×20 and ×100). Dilution played a role in AFFF sonolysis efficiency with optimal PFAS sonolysis rate (4.28 μmol L-1 min−1) at 20 × dilution. Overall AFFF was effectively modelled with a synthetic PFOS solution, attributed to limited matrix effects in AFFF sonolysis and high PFAS concentration (0.18–1.83 g L-1) dominated by PFOS (0.15 – 1.53 g L-1).
Read full abstract