Characterization of Emissions during the Thermal Destruction of Per- and Polyfluoroalkyl Substances

Abstract Description: Some studies have reported destruction and removal efficiencies (DREs) of >99.99% during the thermal destruction of PFAS during the thermal reactivation of spent granular activated carbon (GAC). However, using DRE alone to assess PFAS destruction is insufficient because it does not account for the presence of products of incomplete destruction (PIDs) in gas-phase emissions. This study aimed to characterize emissions during the thermal destruction of adsorbed and non-adsorbed PFAS. Experiments were conducted with pure PFAS standards, pure PFAS standards adsorbed to GAC, PFAS/hydroxide mixtures, and PFAS/natural organic matter (NOM) mixtures in a thermogravimetric analyzer (TGA) furnace under pyrolysis conditions to characterize and (semi-)quantitatively measure PIDs. Off-gas from thermal destruction experiments were collected with a sampling train configuration adapted from United States Environmental Protection Agency’s (EPA’s) Other Test Methods (OTM) 45 and 50. The sampling train consisted of impingers to trap gaseous compounds soluble in water amended with 0.1 M NaOH and a SUMMA canister to capture volatile fluorinated compounds that pass through the impingers. In addition, semi-volatile PIDs were captured with an XAD resin trap. Impinger solutions and TGA pan residues were analyzed for fluoride and other anions and cations using ion chromatography. Targeted PFAS analysis of the transfer line rinses and impinger solutions was performed by liquid chromatography-mass spectrometry (LC-MS). Targeted and non-targeted analysis of volatile and semi-volatile fluorinated compounds was conducted by gas chromatography-mass spectrometry (GC-MS). Results show that thermal destruction of PFAS can result in the formation of polar PIDs as well as non-polar semi-volatile and volatile PIDs. Apart from compounds targeted by OTMs 45 and 50, non-targeted suspect screening methods led to the identification of additional PIDs, including two isomers each of octafluorobutene (C4F8) and perfluoroheptene (C7F14). Using a surrogate compound approach, semiquantitative analysis of C4F8 isomers suggested their contribution was up to 70% to the total fluorine mass balance for experiments with conducted with perfluorobutane sulfonic acid (PFBS). The formation of PIDs decreased in the following order: non-adsorbed/no addition (~70%) < non-adsorbed/Ca(OH)2 (~49%) < non-adsorbed/NaOH (~27%) < non-adsorbed/NOM (~20%) < adsorbed/dry GAC (~4%). During this research, PIDs targeted by standard EPA methods as well as novel PIDs were identified. Their behavior in thermal oxidizers (afterburners) used for air pollution control needs to be understood in future research to identify operating conditions that ensure their destruction.