Distinguishing PFAS biotransformation from defluorination and destruction: a fluorine-fate evidence framework

Per-and polyfluoroalkyl substances (PFAS) are a structurally diverse class of persistent, mobile, and bioaccumulative contaminants, with more than one megatonne produced annually and continuing releases throughout manufacture, use, and disposal (Evich et al., 2022). Volatile PFAS and their precursors can undergo long-range atmospheric transport and transform into persistent terminal PFAS, facilitating their circulation among air, water, soil, sediment, and biota (Renella and Quintarelli, 2026;Evich et al., 2022). Their environmental recalcitrance is strongly associated with the exceptional stability of the C-F bond, for which reported bond dissociation energies depend on the molecular environment but are generally high in highly fluorinated organic compounds (approximately 485 kJ mol⁻¹ as a representative value), together with its short bond length, high electronegativity, chemical stability, and limited biological accessibility (Khan, 2025;Renella and Quintarelli, 2026). PFAS are consequently detected from ng L⁻¹ to µg L⁻¹ across environmental matrices, while contaminated agricultural and industrial soils can remain long-term sources to groundwater and food webs (Khan, 2025;Khan, 2026a;Kareem and Khan, 2026). These physicochemical characteristics underpin PFAS persistence and provide the chemical basis for distinguishing molecular transformation from C-F bond cleavage and the extent of fluorine and carbon conversion associated with destructive treatment (Figure 1A).This persistence has stimulated extensive research into PFAS remediation, including adsorption, ion exchange, membrane separation, foam fractionation, soil stabilisation, thermal treatment, electrochemical processes, and plasma technologies. Separation and stabilisation can substantially decrease aqueous concentrations or suppress leaching, including reported reductions of approximately 60 to 90% following soil stabilisation, but generally transfer or immobilise rather than destroy the fluorinated mass (Uddin et al., 2026). Destructive technologies are therefore increasingly evaluated using defluorination and mineralisation rather than parent-PFAS removal alone (Tshangana et al., 2025;Meegoda et al., 2022;Wen et al., 2026). These non-biological technologies provide useful comparators for evaluating the relationship between parent-PFAS degradation and fluorine removal; they are included here to clarify the evidence required to substantiate destructive remediation, rather than to imply biological mechanisms. Biological approaches have attracted particular interest because bacteria, fungi, algae, consortia, and enzymes can operate under relatively mild conditions, with proposed involvement of fluoroacetate dehalogenases, cytochrome P450s, reductive dehalogenases, oxygenases, hydrolases, peroxidases, and laccases (Liang and Zhao, 2026;Khan et al., 2023a). Metagenomics, CRISPR, multi-omics, metabolic modelling, synthetic biology, and enzyme engineering are further accelerating the search for biological PFAS transformation and biodefluorination pathways (Khan, 2025;Olawade et al., 2025;Che et al., 2025).The concern is that terminology has often advanced faster than the evidence. Parent-PFAS disappearance can result from adsorption, bioaccumulation, sequestration, precipitation, volatilisation, or phase redistribution, while transformation may simply redistribute fluorine among shorter-chain PFAS, fluorotelomer products, or other unmonitored organofluorines (Cao et al., 2025;Huang et al., 2025). Even genuine molecular transformation therefore does not necessarily constitute destruction. For example, Labrys portucalensis F11 removed approximately 90% PFOS and 58% 5:3 FTCA after 100 days, with PFOS removal reaching 96% after 194 days, but only 21% of 6:2 FTS was removed after 100 days. PFOS transformation generated PFHpS, PFHxS, PFHxA, PFPeA, PFBA, PFPrA, and three defluorinated products, whereas 5:3 FTCA produced shorter-chain and unsaturated fluorotelomer products and 6:2 FTS produced 4:2 FTS (Wijayahena et al., 2025). These results demonstrate biological transformation and partial defluorination, but not complete destruction. Accordingly, the evidence needed to characterise PFAS remediation should distinguish several related but nonequivalent outcomes: parent-PFAS disappearance, transformation-product formation, direct evidence of C-F bond cleavage, fluoride release, fluorine mass balance, and carbon mineralisation (Figure 1B). These outcomes form an evidence framework rather than a universal sequence: transformation may occur without demonstrable C-F bond cleavage; defluorination may occur without complete carbon mineralisation; and carbon-chain transformation may occur without complete removal of organofluorine. Importantly, these outcomes should not be treated as interchangeable: biological transformation may occur without demonstrable C-F bond cleavage, whereas biodefluorination specifically requires evidence of fluorine removal from the organic molecular framework. Extensive defluorination and carbon mineralisation should likewise be evaluated as distinct endpoints, because evidence for one does not, by itself, establish the other. This Opinion therefore argues that parent-PFAS disappearance alone cannot establish biodegradation or destructive remediation. Instead, the terminology should reflect the demonstrated outcome, with convincing fluorine-fate evidence required to substantiate claims of biodefluorination and PFAS destruction.A scientifically meaningful definition of PFAS biodegradation should distinguish removal, transformation, defluorination, and mineralisation according to the evidence obtained. Removal indicates a concentration decrease, while sorption or sequestration represents redistribution without destruction (Wackett, 2021;Cao et al., 2025). PFAS biotransformation refers to biological conversion of the parent compound into one or more products, irrespective of whether C-F bond cleavage occurs. Partial biodegradation may involve biological structural modification while retaining substantial organofluorine. Biodefluorination specifically requires evidence of C-F bond cleavage and fluorine removal from the organic molecular framework. Mineralisation involves conversion of organic constituents towards inorganic end products and must be distinguished from defluorination, because fluoride release alone does not demonstrate mineralisation of the carbon skeleton (Wackett, 2021;Wackett and Robinson, 2024). Claims of PFAS destruction should therefore be supported by convincing evidence of defluorination and an assessment of fluorine fate, while claims of mineralisation additionally require evidence of carbon-product conversion and, where feasible, carbon mass balance (Wackett and Robinson, 2024;Wen et al., 2026). Accordingly, claims of mineralisation require evidence of carbon-product conversion and, where feasible, carbon mass balance. These outcomes depend on PFAS structure, redox conditions, oxygen availability, enzyme specificity, and microbial interactions (Skinner et al., 2025;Che et al., 2025;Renella and Quintarelli, 2026).The high stability of the C-F bond makes its biological activation challenging, although microbial C-F cleavage has been documented across fluorinated aliphatic and aromatic compounds and increasingly in PFAS-relevant systems (Natarajan et al., 2005;Liu et al., 2024). The distinction between transformation and defluorination is illustrated by Cunninghamella elegans, which completely consumed 100 mg L⁻¹ 6:2 FTOH within 48 h while accumulating 5:3 FTCA. The product was identified using ¹⁹F NMR, GC-MS, and LC-MS, and inhibition by 1-aminobenzotriazole implicated cytochrome P450 activity; 5:3 FTCA itself inhibited P450-mediated metabolism, contributing to its accumulation (Khan and Murphy, 2023). Similarly, individual photocatalytic and fungal treatments of 100 ppm PFOA achieved approximately 35 to 40% degradation but only 20 to 30% defluorination, whereas sequential photocatalysis-fungal treatment increased degradation to 90% and defluorination to 60%; photocatalysis produced short-chain PFCAs while fungal treatment predominantly produced 5:3 FTCA (Khan et al., 2023b). This combined treatment provides a direct comparison of physicochemical and biological processes, illustrating how complementary treatment steps can improve both transformation and defluorination. These results show that parent-compound conversion and fluorine removal can diverge substantially, even when transformation is extensive.Precursor transformation provides an additional warning. N-EtFOSAA can biologically transform to PFOS, meaning that disappearance of a precursor can generate rather than eliminate a persistent terminal PFAS (Thapa et al., 2026). Conversely, some structures can undergo transformation that initiates genuine defluorination. Jin et al. (2023) showed that susceptible ether PFAS containing a CH₂ group adjacent to an ether linkage or a C=C bond can undergo oxidative or hydrolytic O-dealkylation to unstable fluoroalcohols that subsequently defluorinate. Transformation-product identification is therefore essential for resolving PFAS fate, and reaction libraries can support systematic prediction and interpretation of environmental and biological pathways (Weber et al., 2022). These examples show that transformation-product profiles are mechanistically informative, but their implications for fluorine fate must be established independently.Direct measurement of fluoride provides stronger evidence of C-F bond cleavage because hydrolytic, reductive, and eliminative pathways can release F⁻ alongside organic products (Simon et al., 2025). Yu et al. (2020) demonstrated microbial reductive defluorination of two C6 per-and polyfluorinated compounds by an organohalide-respiring community, coupling product formation with direct F⁻ release under lactate-fed conditions; although Dehalococcoides was present, it was not responsible for the observed activity. More recent pH-based screening has expanded high-throughput discovery of defluorinases, but proton release remains an indirect indicator compared with direct fluoride quantification (Simon et al., 2025). Fluoride yields should be reported relative to the theoretical fluorine content of the substrate, with corrections for background fluoride, analytical recovery, and substrate stoichiometry.The fungal transformation of PFOA to 5:3 fluorotelomer carboxylic acid (5:3 FTCA) has been experimentally reported, with 5:3 FTCA identified as the main product of fungal treatment (Khan et al., 2023). PFOA contains 15 F atoms, whereas 5:3 FTCA contains 11; therefore, a hypothetical 1:1 molar conversion corresponds to the potential release of four F atoms per transformed PFOA molecule.Figure 2 uses this experimentally reported transformation as the basis for an illustrative fluorine-fate calculation, assuming 60% conversion. The assumed conversion and calculated fluorine distribution are hypothetical and should not be interpreted as experimentally measured outcomes. The calculation illustrates how parent transformation may retain most substrate fluorine in an organic product; it does not establish that this fluoride yield or product distribution occurred experimentally.Non-biological treatment studies provide a useful comparator for the same analytical distinction. Co(II) phthalocyanine with Ti(III) citrate achieved 10.9% defluorination of technical PFOS and 54.5% of branched PFOS isomers after 8 h, whereas vitamin B12 achieved only 2.4% defluorination (Mirabediny et al., 2024). Similarly, cold atmospheric plasma produced 99.99% PFOS degradation but only 35% defluorination, with stepwise transformation products and fluorine mass-balance recoveries of 31-106% (Kumar et al., 2026). Although these systems are not evidence of microbial activity, they reinforce the need to quantify degradation and defluorination separately, a principle equally relevant to biological remediation.A fluorine mass balance should link the initial fluorine inventory to released F⁻ and residual fluorine in parent PFAS, identified and unidentified transformation products, biomass, particulate phases, and other relevant compartments. Targeted LC-MS/MS should be complemented by TOF, EOF, AOF, HRMS, nontargeted analysis, and transformation-product libraries to capture organofluorine beyond the targeted analyte list (Forster et al., 2023;Idjaton et al., 2024;Liu et al., 2025). TOF, EOF, and AOF represent operationally defined fluorine pools determined by different analytical procedures. Their measured values are method-dependent and should not be treated as directly interchangeable or summed, because the pools may overlap and their recoveries may differ. Mass-balance closure should be interpreted in light of analytical recovery, measurement uncertainty, and potentially unmeasured fluorine pools. Carbon-product analysis is additionally required to support mineralisation claims.The literature demonstrates that microorganisms can transform selected PFAS and that microbial C-F cleavage is chemically achievable, although the strength of evidence varies substantially. A metaanalysis of 97 microbial PFAS biotransformation studies found that transformation was more likely under aerobic conditions, in defined or axenic cultures, at higher PFAS concentrations, and for compounds containing fewer fluorine atoms. Activity also depended on chain length, branching, and headgroup. Anaerobic systems with defined electron donors and acceptors, identified organisms or enzymes, and comprehensive transformation-product data remain comparatively underrepresented (Skinner et al., 2025). Cross-study comparisons must therefore account for both the reported endpoint and experimental conditions. Results from defined systems or elevated PFAS concentrations may not translate directly to environmental settings, where matrix properties affect partitioning, bioavailability, microbial activity, and analytical recovery.Parent depletion is the most frequently reported observation, but microbial systems may combine adsorption, accumulation, transformation, and abiotic redistribution. Saifur et al. (2026) reported 10.9-38.3% adsorption without parent-PFAS transformation in anaerobic wastewater cultures over 90 days.In aerobic cultures after 21 days, adsorption ranged from 26.9 ± 1.2% to 55.8 ± 1.4%, and biotic accumulation was 13.35-17.55%. Transformation was reported for PFOA (47.87 ± 1.6%), PFOS (28.96 ± 0.6%), and 6:2 FTS (43.1 ± 1.0%), with shorter-chain products and limited defluorination. These concurrent processes show why single-phase measurements may not represent the overall PFAS mass balance. Representative studies and their evidence limitations are summarised in Table 1.Matrix effects are critical in wastewater, biosolids, soil, and sediment. In wastewater and biosolids, partitioning between water and solids can make apparent aqueous removal difficult to distinguish from transfer to sludge; biosolids composition and microbial weathering may also alter PFAS partitioning over time (Lewis et al., 2023). In soil, retention and leaching depend on soil properties and application conditions. Field-relevant lysimeter experiments have demonstrated differences in short-and longchain PFAS distribution between soil and leachate following biosolids application (Peter et al., 2024), while the delivery matrix can influence PFAS mobility and bioavailability (Sidhu et al., 2026). Studies should therefore assess aqueous and solid-associated PFAS, where relevant, and account for changing partitioning during incubation. Groundwater assessments should also consider sorption to aquifer materials and the accumulation of transformation products.Transformation-product analysis provides stronger mechanistic evidence but remains incomplete when restricted to conventional target lists. Liu et al. (2025) (Yang et al., 2026). Background-corrected fluoride release and limited transformation in controls supported reductive microbial transformation and defluorination, but the long-term product profile demonstrates why removal alone is This illustrates the of beyond defined aqueous although further is needed across groundwater and evidence should be under relevant conditions, with and analytical to depletion and product identification establish initial while direct fluoride measurement and fluorine-fate analysis are needed to quantify defluorination. Carbon-product analysis the of and must also be evaluated as distinct outcomes. Although fluoride release provides evidence of C-F bond cleavage, it does not demonstrate that the products are or requires assessment of transformation-product and using chemical and biological et al., et al., 2026). studies should assess relevant aqueous and phases, while groundwater studies should account for background fluoride and potential sorption to aquifer Fluoride and closure should be reported alongside analytical and rather than assumed from results in defined studies remain even when defluorination has not been as they can susceptible transformation and The research is transformation-product formation to fluorine fate at relevant concentrations and in studies should experiments that establish and support mass-balance evidence should directly biological engineering than organisms for PFAS or parent-compound disappearance, remediation systems should target the whether biological transformation, C-F bond cleavage, extensive defluorination, or This requires of microbial with enzyme metabolic and The should be to distinguish systems that transform PFAS from that while evaluating carbon fate when mineralisation is Importantly, the engineering in their of some biological systems have demonstrated PFAS transformation or defluorination, whereas several proposed approaches remain and require experimental engineering approaches including metabolic synthetic biology, metabolic analysis, and that potentially be to microbial systems of PFAS transformation or defluorination. these approaches should be as engineering has been experimentally demonstrated et al., 2025). In the application of CRISPR, or to other biological systems does not, by itself, establish their for engineering of C-F in selected fluorinated compounds provide potential for enzyme discovery and enzyme associated with hydrolases, and oxygenases, although a organofluorine substrate cannot be assumed to to structurally diverse PFAS (Natarajan et al., 2025). PFAS persistence both chemical stability and biological of is a potential but its to relevant PFAS requires direct (Wackett, identified microbial and and approaches to transfer and defluorination activity, to prediction for et al., 2025). Similarly, a from defluorinated and while activity, the potential of engineering for C-F cleavage (Khan et al., 2026). These establish relevant for defluorination and enzyme but they should not be interpreted as evidence of PFAS defluorination PFAS has been these support a from enzyme discovery with PFAS substrate activity, product and fluoride release as essential structural also the of reaction rather than on a or combine reductive, hydrolytic, and to structures that may be to individual et al., et al., 2025). the proposed of these into a remains an engineering the of individual must be established including assessment of accumulation and fluorine sequential photocatalysis-fungal treatment of PFOA illustrates the potential of physicochemical and biological processes, whereas the provides a biological of long-term microbial transformation and defluorination (Khan et al., et al., 2026). These examples provide evidence for treatment and biological but not establish the of microbial or treatment as a for PFAS remediation. systems be into treatment in which PFAS, physicochemical processes and selected microbial or systems are to transform The and overall of these treatment require using fluorine mass and relevant environmental generated during C-F cleavage can microbial and activity, of fluoride sequestration, and product alongside (Wackett, and Robinson, 2024). remains by matrix microbial stability, and and while systems additionally require and (Tshangana et al., et al., 2025;Olawade et al., 2025). the engineering endpoint should be PFAS transformation or as to the with claims of destructive remediation supported by fluorine-fate rather than parent-PFAS removal under conditions. The distinction between demonstrated and proposed engineering should be when studies and for must be by of distinguishing genuine defluorination from parent-PFAS disappearance, redistribution, or incomplete should quantify parent PFAS and transformation products, fluoride release, and assess fluorine mass balance. et al. (2025) provide a basis for defluorination from fluoride released relative to PFAS removed and the fluorine content of the Targeted and nontargeted analysis, together with TOF, EOF, and AOF should be because targeted may transformation products and pools (Forster et al., 2023;Idjaton et al., TOF, EOF, and AOF are operationally defined measurements of potentially fluorine their results should not be treated as directly interchangeable or A fluorine mass balance should where feasible, for released residual parent PFAS, identified transformation products, and other pools (Kumar et al., et al., 2026). The of mass-balance closure should be interpreted in light of analytical recovery, measurement uncertainty, and the of unmeasured or mineralisation carbon-product analysis must establish conversion of carbon to inorganic as or inorganic rather than its persistence in organic feasible, PFAS can distinguish inorganic carbon from background The persistence and of transformation products should also be evaluated because shorter-chain PFAS and other can remain persistent and or distinct properties et al., et al., and Khan, 2026). and controls are required to distinguish biological from abiotic adsorption, precipitation, and phase redistribution, while and long-term evidence for and PFAS These measurements should be reported alongside experimental conditions, analytical and that systems can be compared on a basis et al., et al., 2026). For proposed this should be at the of that evidence of PFAS transformation or defluorination is established of in parent-compound systems should therefore be evaluated the they rather than a of transformation products, fluoride release, and carbon conversion provide complementary but should be treated as an even extensive defluorination may yield transformation products that remain persistent, mobile, or Accordingly, claims of require direct assessment of transformation-product and cannot be from fluoride release, parent-PFAS or carbon mineralisation alone et al., 2026). This engineering with chemical outcomes and the of systems that PFAS from but persistent organofluorine also provides a distinction between and engineering that claims of are to the because it remediation and the biological selected for A as PFAS may imply substantial environmental even when most fluorine remains in transformation The distinction between and destruction is in PFAS treatment and should be equally to biological studies et al., 2022). The is therefore not whether a can biological but whether reported terminology the measured This distinction is because PFAS strongly on chain length, branching, ether precursor structure, and environmental (Skinner et al., 2025;Che et al., biodegradation claims can discovery and engineering by organisms that PFAS from sorption or precursor transformation rather than C-F directly for fluoride release the required for destructive remediation, while high-throughput defluorination and enzyme engineering provide to and biological systems (Simon et al., et al., 2025). Accordingly, the engineering should from parent-PFAS removal towards defluorination, fluorine-fate and of persistent is essential to support this A framework to PFAS approaches can PFAS fate, chemical and analytical evidence et al., 2024). For biodegradation this should parent transformation products, fluoride release, residual analytical recovery, and, where relevant, carbon treatment need not complete PFAS structural and microbial systems may be most within treatment separation, physicochemical transformation, and biological of susceptible et al., et al., et al., 2025). The same mechanistic support and by fluorinated structures more to biological C-F cleavage et al., 2025). The proposed framework is to make claims more not to studies that demonstrate biotransformation without defluorination. studies remain for and enzymes, but their environmental on the and of the products The should consequently from parent-compound disappearance alone towards supported assessments of PFAS removal, and and and that support was for this or its under conditions. fluorotelomer fluorotelomer fluorotelomer PFAS, per-and polyfluoroalkyl PFOS, indicates that the relevant evidence was not in the for this it does not establish that the not removal, defluorination and mineralisation are distinct outcomes. Fluoride release defluorination, but claims of extensive fluorine removal require a fluorine mass balance. Fluoride release alone does not demonstrate carbon