Before discussing the working performance of our sorbent for PFAS removal, we briefly introduce the design of our perfluoropolyether-containing ion-exchange resin (PFPE-IEX+). Firstly, the inclusion of PFPE segments provides fluorine-fluorine hydrophobic interactions with the perfluoroalkyl segments of PFAS, and such interactions have been shown to be rapid and selective according to previous studies13,17. Secondly, compared with the traditional monomers such as acrylate or methacrylate, the use of styrene-modified monomers not only increases the stability of the sorbent in aqueous solutions under different pHs, but also increases the hydrophobicity of the sorbent, potentially further increasing the capacity for sorption of PFAS18. In addition, the incorporation of quaternized ammonium groups further enhances the capture of anionic PFAS in aqueous solutions via extra electrostatic attraction. The typical chemical structure and synthetic routine of PFPE-IEX + are shown in Fig. 1a.
PFPE-IEX + resin design and synthesis
Styrenically functionalized perfluoropolyethers (sPFPE) were firstly synthesized through previously reported alkylation reaction between commercially available hydroxyl-terminated perfluoropolyether PFPE-OH (molecular weight Mn ~ 2000 Da) and 4-vinylbenzyl chloride (VBC)19. Both 1H and 19F NMR spectra of sPFPE confirm the successful synthesis of the sPFPE monomer, as all the peaks have been successfully assigned, and are shown in Fig. 1b and Supplementary Fig. 1. PFPE-IEX was then synthesized using sPFPE (42 wt%), VBC (48 wt%) and divinylbenzene (DVB, 10 wt%) in the presence of azobis(isobutyronitrile) (AIBN) as the initiator. After purification, the successful synthesis of PFPE-IEX was confirmed by performing solid-state (SS) 13C NMR and Fourier-transform infrared spectroscopy (FTIR). The SS 13C NMR spectrum exhibited two broad signals located at approximate 143 and 125 ppm (peaks a and b, Fig. 1c upper), which can be assigned to the carbon atoms belonging to aromatic groups. Additionally, two peaks at ~ 44 and ~ 37 ppm can be also observed, originating from the alkyl carbons in the polymer backbone (peaks c and d, Fig. 1c upper). Quaternization of PFPE-IEX was conducted using trimethylamine to produce cationic PFPE-containing sorbent PFPE-IEX+. The successful quaternization was again confirmed by SS 13C NMR (Fig. 1c) and FTIR (Fig. 1d). In the SS 13C NMR spectrum, the carbon atoms belonging to the -CH2-Cl group of PFPE-IEX shift from ~ 40 (peak c) to 50 ppm (peak e) after quaternization20,21. Additionally, the newly formed peak f at 65 ppm after quaternization is due to the carbon atom from the -N-(CH3)3 group. FTIR further confirms the successful quaternization of the PFPE-IEX resin, with typical absorption peaks at 3005 cm− 1, 1474 cm− 1, and 950 cm− 1 appearing after quaternization, assigned to the C-H bending and stretching of trimethyl group from quaternized ammonium cations (Fig. 1d)22,23. The above results demonstrate the successful preparation of PFPE-IEX and subsequent quaternization of the sorbent to produce PFPE-IEX+. After being fully dried under high vacuum, the prepared sorbent was ground and passed through sieves to collect powders with desire size of 80–100 µm (Fig. 1e). The swelling ratio is an important additional consideration to confirm the availability of sorbent to achieve long-term continuous PFAS removal in cartridges. Out of the three prepared sorbents (Supplementary Table 1), each possessing varying degrees of crosslink density (4, 10 and 20%), the 10% sorbent demonstrated a favorable swelling ratio of 1.06, indicating a moderate level of swelling (Supplementary Fig. 2). Consequently, we have opted to utilize this 10% PFPE-IEX + sorbent for the subsequent tests.
Excellent stability of PFPE-IEX + was demonstrated under incubation in both acidic and alkaline environments. By respectively immersing the resin in acidic (pH = 2.0) and alkaline aqueous solution (pH = 10.0) at elevated temperature 60°C, the sorbent was collected at pre-determined time points up to a period of 56 days, followed by centrifugation and dried for weight measurements, FTIR, SS 13C NMR and 19F NMR analyses. The results in Supplementary Fig. 3 demonstrate that no significant mass loss over the 56 days under both conditions was observed, indicating good stability of PFPE-IEX+. FTIR and SS 13C NMR spectra in Supplementary Fig. 4 and Fig. 5 clearly show that no changes in structure were observed when comparing the spectrum of PFPE-IEX + before and after immersion and incubation in aqueous solutions under two different pH conditions. Furthermore, 19F NMR spectroscopy was employed to assess the residual fluorine in the supernatant following centrifugation and no fluorine signal was detected under the current testing conditions (Supplementary Fig. 6). In summary, the polystyrene matrix of PFPE-IEX + shows good stability in both acidic and alkaline conditions, minimising hydrolysis in aqueous solutions. In addition, small-angle X-ray scattering (SAXS) experiment on the swollen PFPE-IEX + reveals a broad peak at approximately q = 0.07 Å−1 (Supplementary Fig. 7). This indicates the presence of aggregate clusters, likely originating from the PFPE segments, potentially further contributing to PFAS capture.
PFPE-IEX + resin characterization
Equilibrium sorption
The removal efficiency of 11 different PFAS, including short-chain perfluoroalkyl carboxylic acids (PFCAs, CnF2n+1COOH, n < 7) and perfluoroalkyl sulfonic acids (PFSAs, CnF2n+1SO3H, n < 6), long-chain PFCA (n ≥ 7) and PFSA (n ≥ 6)24,25, and one major emerging PFAS, ammonium salt of hexafluoropropylene oxide dimer acid (HFPO-DA) was tested using PFPE-IEX + at an initial concentration of 100 µg/L (ppb) for each PFAS. Two commercially available sorbents were used as controls, including granular activated carbon (GAC, 20– 40 mesh particle size) and anion exchange resin (IEX, Amberlite IRA-410, 20–25 mesh particle size). The addition of 20 mg/L (ppm) of humic acid and 200 mg/L of sodium chloride (NaCl) was to simulate the natural PFAS contaminated environment. After treatment for 24 h, > 99% removal efficiency for all tested PFAS (> 98% for PFBA) was achieved as quantified using liquid chromatography with tandem mass spectrometry (LC-MS/MS), highlighting the superior removal efficiency of PFPE-IEX + for capturing PFAS compared with commercially available GAC and IEX sorbents (Fig. 2b). To be more specific, PFBA is considered as one of the most challenging PFAS to be removed from the environment, due to its short chain length and high-water solubility. PFPE-IEX + can efficiently remove > 98% of PFBA, a level much higher compared with GAC and IEX (approximate 80% and 40%, respectively). Such superior performance in sorption efficiency is also reflected in the removal of other types of PFAS, including HFPO-DA. PFPE-IEX + shows > 99.5% removal of HFPO-DA after 24 h of incubation, which is significantly higher compared to that treated by GAC and IEX (83.5% and 35%, respectively).
Sorption kinetics
The sorption kinetics for capturing various PFAS in Milli-Q water were subsequently investigated. 11 types of PFAS were spiked into Milli-Q water, creating an initial concentration of 10 ppb for each PFAS. The results presented in Fig. 2c demonstrate that all 11 PFAS could efficiently removed (> 99%) in less than 30 s, highlighting the rapid recognition and outstanding removal of PFAS compounds using PFPE-IEX + sorbent. It should be noted that no desorption occurred throughout the entire 120 mins sorption period, indicating the robust binding between PFAS and PFPE-IEX + via fluorous and hydrophobic interactions and electrostatic attraction. The findings correspond well with the work previously reported by our group26. We were able to achieve rapid sorption of HFPO-DA at an initial concentration of 100 ppb within 30 s, and similarly, no desorption was observed after 24 h.
Sorption capacity
Sorption isotherm of PFAS using PFPE-IEX + resin was performed to understand the binding mechanisms between PFAS and PFPE-IEX+. To be more specific, HFPO-DA solutions with varied concentrations ranging from 0.1 to 50 ppm were treated with PFPE-IEX + at a fixed concentration of 0.1 mg/mL for 24 h in Milli-Q water. As shown in Fig. 2d, the sorption data was fitted by the Langmuir and Freundlich models using Eq. (1) and Eq. (2), respectively27–29.
$$\frac{{\text{C}}_{\text{e}}}{{\text{Q}}_{\text{e}}}=\frac{1}{{\text{Q}}_{\text{m}}{\text{K}}_{\text{L}}}+\frac{{\text{C}}_{\text{e}}}{{\text{Q}}_{\text{m}}}$$
1
$${\text{l}\text{n}\text{Q}}_{\text{e}}=\text{l}\text{n}{\text{K}}_{\text{F}}+\frac{1}{\text{n}}{\text{l}\text{n}\text{C}}_{\text{e}}$$
2
The residual concentration of HFPO-DA at equilibrium is represented by Ce (mg L− 1), while Qe (mg g− 1) is the amount of HFPO-DA that the sorbent binds at equilibrium. Qm (mg g− 1) is the maximum sorption capacity, KL (L mg− 1) is the Langmuir equilibrium constant representing binding affinity30, while KF ((mg g− 1)(L mg− 1)1/n) is the Freundlich constant and n is the intensity of sorption.
The isotherm parameters for both models were calculated using Equations (1) and (2) and are listed in Table 1. The Langmuir model has a better fit than the Freundlich model with a higher R2 value (0.9233 vs. 0.7793). The calculated KL and Qm are 1.93×105 M− 1 and 577.4 mg/g, respectively, indicating a strong binding affinity between HFPO-DA and PFPE-IEX + sorbent. This further supports the observations in the above section, specifically the absence of desorption within the tested time period. The sorption capacity of HFPO-DA at 577.4 mg/g is significantly higher compared to that of previously reported sorbents under equivalent testing conditions (e.g. crosslinked β-cyclodextrin-containing (β-CD) polymer (KL = 8.8×104 M− 1, Qm = 222.0 mg/g)31, chitosan (CS) modified adsorbent (Qm = 364.6 mg/g)32, fluorogels (KL = 5.9×106 M− 1, Qm = 278.0 mg/g)33, fluorinated hydrogels (Qm = 34.2 mg/g)34 and covalent organic frameworks (KL = 6.3×104 M− 1, Qm = 200.0 mg/g)10). The study demonstrats that our sorbent PFPE-IEX + has a high binding affinity and superior sorption capacity compared with other sorbents in the removal of HFPO-DA.
Table 1
Langmuir and Freundlich constants for the sorption of HFPO-DA using PFPE-IEX+.
Langmuir | Freundlich |
KL (M− 1) | Qm (mg/g) | R2 | KF (mg/g)(L/mg)1/n | n | R2 |
1.93×105 | 577.4 | 0.9233 | 160.6 | 1.458 | 0.7793 |
The maximum sorption capacity of PFPE-IEX + for the removal of seven other types of PFAS, i.e. perfluorobutanoic acid (PFBA), perfluoro-n-pentanoic acid (PFPeA), perfluorohexanoic acid (PFHxA), perfluoroheptanoic acid (PFHpA), perfluorooctanoic acid (PFOA), perfluorobutanesulfonic acid (PFBS) and perfluorooctane sulfonate (PFOS), were also tested in Milli-Q water. Experiments were performed by treating each of the individual PFAS solution at an initial concentration of 0.4 mg/mL using PFPE-IEX + at 0.2 mg/mL in Milli-Q water for 24 h. The results presented in Fig. 2e and Supplementary Fig. 8 indicate the sorption capacities for all tested PFAS using PFPE-IEX + are significantly higher compared with previously reported sorbents. Significantly, PFBA shows a maximum capacity of 524.1 mg/g using PFPE-IEX+, superior to that of previously reported value of 27.89 mg/g by Hossain et.al35 and 121 mg/g by Li et.al36. Also, PFBS as another example, has an sorption capacity at 928.6 mg/g which is much higher than that reported by Li et al. (319 mg/g)36 and Kebria et al. (305 ± 45 mg/g)37. An increase in sorption capacity was observed with increasing PFAS chain length (i.e. C4 PFBA 524.1 mg/g, C5 PFPeA 700.1 mg/g, C6 PFHxA 844.3 mg/g, C7 PFHpA 967.8 mg/g, and C8 PFOA 1178.5 mg/g). This is mainly due to the higher water solubility and weaker fluorous interactions for those PFAS having shorter chain lengths, agreeing well with previously reported literature25.
Regeneration and reusability
Regeneration and reusability are key factors in evaluating the practicality of PFAS sorbents and play a vital role in recycling the sorbed PFAS for use as battery electrolyte additives38. HFPO-DA was again chosen as a representative PFAS to demonstrate the regeneration and reusability of PFPE-IEX+. Five cycles that include both sorption and desorption experiments were performed, with sorption/desorption durations of 30 min for each cycle (Fig. 2f and Supplementary Fig. 9). Desorption was performed by replacing the aqueous solution with the same volume of methanolic salt solution (1% NaCl in methanol). As has been reported in previous work13,26,39, addition of organic solvents in the presence of salt can efficiently release the bound PFAS, with organic solvents weakening the fluorous hydrophobic interactions and the salt breaking the electrostatic attraction. Both 19F NMR (Fig. 2f) and LC-MS/MS (Supplementary Fig. 9) were performed for characterization of the sorption/desorption of HFPO-DA by PFPE-IEX+. As has been observed above, efficient removal of HFPO-DA was achieved in the first sorption cycle (> 99% removal efficiency with capacity of 173 and 174 mg/g determined by 19F NMR and LC-MS/MS, respectively), followed by efficient desorption and release of HFPO-DA in methanolic salt solution. However, it was observed that the desorption efficiency of HFPO-DA was not as high as in the sorption step, likely due to the limited solubility of HFPO-DA in methanol at high concentrations. Nevertheless, PFPE-IEX + maintains a sorption efficiency of > 99% over five cycles without a significant loss in sorption capacity, demonstrating that PFPE-IEX + has excellent regeneration and reusability for repeated PFAS removal.
PFAS management applications
Performance evaluation of PFAS removal using real-life contaminated water
The sorption performance using PFPE-IEX + was further tested in PFAS contaminated potable water and landfill leachate, and compared with commercially available GAC. The total organic carbon (TOC) of landfill leachate was 624.3 mg L− 1, while that of potable water was non-detectable (< 0.1 mg L− 1). Metal ion concentrations for both water matrices were quantified using inductively-coupled plasma optical emission spectrometry (ICP-OES) and the results are shown in Supplementary Table 2. Both water matrices were spiked with 11 different PFAS (Fig. 2a), with each PFAS having an initial concentration of low ppb level (1–10 ppb), followed by individual treatment for 24 h using PFPE-IEX + and GAC. The results in Fig. 3a demonstrate that in potable water, both sorbents show highly efficient removal of all PFAS, indicating that both sorbents are above to remove multiple PFAS in potable water i.e. the solution with low levels of co-contaminants. However, the composition of leachate water includes both dissolved organic and inorganic materials, and can largely interfere with the sorption of PFAS at low concentrations by sorbents, resulting in poor removal performance. In this work, after treating the PFAS contaminated leachate using PFPE-IEX + or GAC, the results in Fig. 3b demonstrate that PFPE-IEX + shows significantly higher removal of all of the 10 PFAS tested compared with the commercially available GAC. For example, after treatment by PFPE-IEX+, high removal efficiency of the majority of the PFAS species was still maintained, with > 99% removal for short chain PFAS including PFHxA, PFHpA and PFBS, > 98% removal of all long chain PFAS tested and > 99% removal of HFPO-DA being observed. In contrast, GAC shows only 30–40% of the short-chain PFAS and only 40% of HFPO-DA, presumably due to interference from the complex components i.e. co-contaminants present in the leachate. These findings demonstrate that even in the presence of complicated components, the sorbent PFPE-IEX + still possesses outstanding sorption performance for different types of PFAS, with higher sorption capacity and selectivity compared with GAC, and can be potentially applied to removal PFAS from highly contaminated environmental and industrial sources.
Sorption kinetics of the 11 spiked PFAS (Fig. 2a) at an environmentally relevant concentration (2 ppb) by our sorbent PFPE-IEX + at two different concentrations of 0.5 and 5 mg/mL of sorbent were further investigated in leachate. The results shown in Fig. 3c demonstrate that the PFDA and PFOS, the two PFAS with the longest perfluoroalkyl chain lengths in PFCAs and PFSAs respectively among all of the 11 PFAS, were effectively removed (> 99%) within 30 seconds, while for the PFAS with shorter chain lengths, at least 5 min was needed for the sorption to reach equilibrium. For the sorbent at higher concentration i.e. 5 mg/mL, fast removal of the PFAS was observed except for PFBA (Fig. 3d), indicating rapid recognition of the PFAS by PFPE-IEX + without significant interference from the other components of the leachate.
Continuous PFAS removal and regeneration in a cartridge
Encouraged by the above excellent performance of PFPE-IEX + in the removal of PFAS from real contaminated water, we then proceeded to develop a cartridge for removal of PFAS in a continuous process. The empty cartridge was equipped with a screw cap and two frits (inlet and outlet) with nominal pore size of 20 µm to prevent any leakage of the PFPE-IEX + resin (particles of approximate diameter ~ 100 µm). Cartridges filled with approximately 2 grams of GAC or PFPE-IEX + were tested (Fig. 4a). The water flow speed was controlled by a peristaltic pump and set at 3 mL min− 1. The breakthrough curve for the column was determined by plotting the ratio of the [PFAS]t/[PFAS]0 against the elution volume ([PFAS]t and [PFAS]0 are each PFAS concentration of effluent and influent, respectively). The observation of 10% breakthrough can be used as an indication that early breakthrough is observed13,40.
The study compared the effectiveness of two different sorbents in removing PFAS from leachate by measuring PFAS breakthrough during triplicate regeneration cycles. For the PFPE-IEX + sorbent, in the first PFAS removal stage, no breakthrough of PFAS was observed. Even after three removal and regeneration cycles, no PFAS breakthrough was observed for the PFPE-IEX + sorbent. This suggests that the PFPE-IEX + sorbent is highly effective and has a high capacity for PFAS removal. In contrast, the GAC sorbent showed high breakthrough during the first PFAS removal stage. This suggests that the GAC sorbent was not as effective as the PFPE-IEX + sorbent in removing PFAS from the leachate. GAC likely reached its saturation point rapidly during the first regeneration cycle, leading to reduced removal efficiency. This study found that the PFPE-IEX + sorbent was superior to GAC in removing PFAS from leachate, as it exhibited efficient removal without breakthrough even after multiple regeneration cycles. This indicates that the PFPE-IEX + sorbent is a promising option for PFAS removal from leachate in waste management and environmental remediation applications (Fig. 4b).
Subsequently, continuous removal of PFAS from potable water having an initial PFAS concentration at 1–2 ppb was performed and the results are illustrated in Fig. 4c. Approximate 12 litres of potable water were continuously passed through the PFPE-IEX + cartridge for three days. Note that less than 0.1% breakthrough of all types of PFAS from C4 to C10 was observed over the 12 litres of continuous elution, indicating that the cartridge filled with PFPE-IEX + sorbent has outstanding efficiency in the removal of PFAS from potable water. Furthermore, as shown in Fig. 4d, it was possible to achieve approximate 100% recovery of the sorbed PFAS compounds using NaCl containing methanol solution, indicating once again the great efficiency of PFPE-IEX + cartridge in releasing the sorbed PFAS. It should be noted that PFAS with carboxylic acid charged groups (i.e. PFBA, PFPeA, PFHxA, HFPO-DA, PFHpA, PFOA, PFNA, and PFDA) were washed out first during the elution process, while those having a sulfonic acid group (i.e. PFBS, PFHxS, and PFOS) were eluted later (Fig. 4d). Remarkably, PFAS was concentrated from approximately 12 litres of water solution into a 150 mL methanol salt solution, which could be dried and used as electrolyte additives in batteries. These results demonstrate the excellent performance of PFPE-IEX + cartridge in removing PFAS from both leachate and potable water while also facilitating the release of sorbed PFAS into a concentrated solution for potential reuse purposes.
Recycling PFAS as an electrolyte additive in batteries
The recycled PFAS from the cartridges (Re-PFAS) was subsequently introduced into a battery system as an electrolyte additive, aimed at enhancing battery performance. In this study, the ion-conducting salt employed was 2 M zinc sulphate heptahydrate (ZnSO4·7H2O), with water serving as the solvent. Specifically, 0.5 mg/mL of Re-PFAS containing 11 different types of PFAS was incorporated into the electrolyte solvent for zinc battery testing.
In the Zn|Zn symmetric cell containing the Re-PFAS additive, the cell was disassembled following 10 cycles, and the collected zinc electrode underwent characterization using X-ray photoelectron spectroscopy (XPS) to analyse its surface composition. The F 1s profile in Fig. 5a reveals the presence of Zn-F and C-F bonds at the surface of zinc foil. The C-F bonds originate from residual PFAS in the electrolyte, consisting -CF2 and -CF3 bonds. The Zn 2p XPS spectrum confirms the formation of Zn-F bond with peaks identified at 1048.6 eV and 1025.5 eV. These outcomes provide conclusive evidence of the formation of a ZnF2 layer during the cycling process and the concurrent decomposition of PFAS. Scanning electron microscopy (SEM) was employed to perform a comparative analysis of the surface morphology of zinc foils after 10 cycles, both with and without the presence of PFAS, as shown in Fig. 5c and 5d. Notably, the SEM images clearly illustrate that the inclusion of PFAS as an electrolyte additive lead to a significant suppression of dendrite growth on the surface of the zinc electrode.
To further assess how PFAS additives contribute to the enhancement of battery performance, a series of electrochemical tests were carried out. Zn-Zn symmetric cells were assembled and subjected to cycling at a current density of 0.5 mA cm-2 and areal capacity of 0.25 mAh cm-2. As shown in Fig. 5e, the Zn symmetric cell containing PFAS additives exhibited a consistent and stable performance over a span of 400 hours. In contrast, the control cell with the standard electrolyte displayed an initially high overpotential and unstable cycling behavior within the first 50 hours. After approximately 125 hours, the voltage of this symmetric cell experienced a sudden increase, leading to instability over the subsequent 100 hours, ultimately resulting in a short circuit at 187 hours. Zn|Cu half cells were employed to assess the Coulombic Efficiency (CE), as illustrated in Fig. 5f. Following an initial activation period, CE of the control electrolyte exhibits sudden fluctuations, indicative of significant side reactions and the potential for short-circuiting. In contrast, CE of the Re-PFAS cell remains stable over 200 cycles, consistently exceeding 99.6%. This outcome corroborates the earlier analysis, indicating that the inclusion of PFAS in the electrolyte contributes to the stabilization of the Zn stripping and plating processes.
Full cells were also assembled to assess the impact of Re-PFAS on the cycling performance. These cells utilized vanadium oxide (V2O5) as the cathode material, chosen for its high theoretical Zn-storage capacity41, and zinc as the anode. As depicted in Fig. 5g, the initial specific capacity was nearly identical for both the control and Re-PFAS-added electrolytes. However, after 50 cycles, the specific capacity of the battery with the control electrolyte began to decline, ultimately decreasing to 22.6 mAh g-1 after 300 cycles, with a retention rate of approximately 17%. Conversely, the battery with Re-PFAS-added electrolyte exhibited a more stable specific capacity compared to the control electrolyte, maintaining a capacity of 65 mAh g-1 after 300 cycles, with a retention rate of above 50%. The CE data also reveals significant insights. With the control electrolyte, the CE began to fluctuate at 150 cycles and the maximum CE exceeded 100%, suggesting that undesired reactions occurred within the battery. Conversely, the CE remained stable at > 99.6% throughout the cycling process when Re-PFAS was introduced. This stability indicates that the addition of Re-PFAS in the electrolyte contributed to a more consistent performance, and mitigated the undesired reactions compared to the control electrolyte.