Comparative immobilization of 30 PFAS mixtures onto biochar, clay, nanoparticle, and polymer derived engineered adsorbents: Machine learning insights into carbon chain length and removal mechanism
Masud Hassan, Ravi Naidu, Fangjie Qi, Bing Wang, Liang Wang, Asadi Srinivasulu, Amal Kanti Deb, Jianhua Du, Yanju Liu
Abstract
Per- and polyfluoroalkyl substances (PFAS) are a group of fluorinated chemicals that cause potential risk in PFAS-impacted soil and water. The adsorption efficiency of 30 PFAS mixtures using different adsorbents in environmentally relevant concentrations was investigated. Different meso/microporous designed adsorbents (n = 7) were used for PFAS adsorption and their interfacial interactions. The adsorbents were tested for their ability to remove PFAS mixtures, including perfluoroalkyl sulfonic acids (PFSAs, n = 7, C4-C10), perfluoroalkyl carboxylic acids (PFCAs, n = 11, C4-C14), fluorotelomer sulfonic acids (FTSs, n = 4), perfluoroalkane sulfonamido acetic acids (FASAAs, n = 3, C8), perfluoroalkane sulfonamides (FASAs, n = 3, C8) and perfluoroalkane sulfonamidoethanols (FASEs, n = 2, C8). The overall removal rate of 30 PFAS was recorded as 86.20–89.29 %, 87.63–90.33 %, and 67.07–93.61 % for microporous biochar/modified biochar, halloysite nanoclays, and mesoporous polymer composites-based adsorbents, respectively. The presence of sugarcane bagasse-derived biochar, iron nanoparticles, and β-cyclodextrin in the composite adsorbents enhances the sorption of PFAS. Higher adsorption efficiency was observed for long-chain PFCAs, PFSAs, FTSs, FASAAs, FASAs, and FASEs, whereas, complete removal of short-chain PFCAs, PFSAs, and FTSs is still challenging by using all the studied adsorbents. The carbon chain length and head groups of PFAS play a vital role in removing PFAS. The correlation coefficient (R 2 ) values between removal rate and carbon chain length, for PFCAs (n = 11), and PFSAs (n = 7) were found as 0.73, and 0.31 respectively. Appropriate machine learning tools including efficient linear least squares, Gaussian process regression, and stepwise linear regression, were applied to fit experimental data and assess model accuracy. • Microporous and mesoporous adsorbents were used to remove the mixture of 30 PFAS. • Carbon chain length and head groups of PFAS play a vital role in the adsorption of PFAS mixtures. • Removal of short-chain PFAS is still challenging compared to long-chain PFAS for all the adsorbents. • Magnetic mesoporous polymer beads (M7) performed better in removing PFAS at environmentally relevant concentrations. • Correlation and ML analysis identified key influential factors and removal mechanisms for separation of PFAS.