Per and Polyfluoroalkyl Substances (PFAS) pose significant health and ecological risks due to their persistence and toxicity. However, knowledge of effective degradation methods remains limited. This review evaluated the effectiveness of computational and experimental approaches for transforming and degrading PFAS. Statistical comparisons and meta-analysis, including Bland-Altman and t-test analyses, were used to assess the occurrence, behaviour, degradation, and modelling of PFAS across different environmental and technospheric compartments. PFAS concentrations ranged from 0. 3 to 1200 ng/L in surface waters and 0. 1–4500 ng/g in sediments. Computational models, including DFT and hybrid ML-MD approaches, predict degradation efficiencies up to 98 % with reaction rates of 1. 2–1. 5 mol/L/min and energy barriers of around 22. 7–30. 1 kcal/mol. Statistical analyses show strong model accuracy (R² = 0. 91) and a mean degradation efficiency of 92 %, while optimised catalysts such as Fe-based catalysts achieve 98 % efficiency and reduce costs to as low as 1. 2/kg. Findings from Bland-Altman and t-test analyses show that computational models outperformed PFOA experiments but underperformed for FTAs and Long-Chain PFAS. This paper concludes that employing computational studies with experimental data promotes the development of more effective and sustainable PFAS degradation strategies. This paper recommends advancing computational tools and interdisciplinary approaches to enhance PFAS remediation techniques.
Ofori et al. (Tue,) studied this question.