Surface Modification of Graphene Oxide for Fast Removal of Per- and Polyfluoroalkyl Substances (PFAS) Mixtures from River Water

Per- and polyfluoroalkyl substances (PFAS) make up a diverse group of industrially derived organic chemicals that are of significant concern due to their detrimental effects on human health and ecosystems. Although other technologies are available for removing PFAS, adsorption remains a viable and effective method. Accordingly, the current study reported a novel type of graphene oxide (GO)-based adsorbent and tested their removal performance toward removing PFAS from water. Among the eight adsorbents tested, GO modified by a cationic surfactant, cetyltrimethylammonium chloride (CTAC), GO-CTAC was found to be the best, showing an almost 100% removal for all 11 PFAS tested. The adsorption kinetics were best described by the pseudo-second-order model, indicating rapid adsorption. The isotherm data were well supported by the Toth model, suggesting that PFAS adsorption onto GO-CTAC involved complex interactions. Detailed characterization using scanning electron microscopy-energy dispersive X-ray spectroscopy, Fourier transform infrared, thermogravimetric analysis, X-ray diffraction, and X-ray photoelectron spectroscopy confirmed the proposed adsorption mechanisms, including electrostatic and hydrophobic interactions. Interestingly, the performance of GO-CTAC was not influenced by the solution pH, ionic strength, or natural organic matter. Furthermore, the removal efficiency of PFAS at almost 100% in river water demonstrated that GO-CTAC could be a suitable adsorbent for capturing PFAS in real surface water.


Text S1. Preparation of reduced graphene oxide (rGO)
The synthesis of rGO was conducted following the technique described by De Silva et al. (De Silva et al., 2018) with some minor adjustments.In summary, a solution of GO at a concentration of 0.1 mg/mL was prepared by subjecting dried GO to sonication in distilled water.The 100 mg of ascorbic acid was added to a 100 mL solution.The pH of the medium was modified to about 10 by introducing an NH4OH solution in order to enhance colloidal stability via electrostatic repulsion (Fernández-Merino et al., 2010).The mixture was stirred at 65 C.The resultant suspensions were subjected to centrifugation at a speed of 4500 revolutions per minute for a duration of 15 min.The solid was gathered and subjected to freezing at a temperature of -20 C for 24 h, after which it underwent freeze-drying at -45 C for 72 h.The resulting composite is denoted as rGO in the subsequent sections of the study.

Text S2. Preparation of ethanolamine (EA) functionalized graphene oxide (GO)
A GO-EA composite was synthesized using the approach described by Ma et al. (Ma et al., 2018) with some minor adjustments.In summary, a 50 mg sample of GO was evenly distributed in 10 mL of water using ultrasonication for 7 h.Subsequently, the solution was put into a 50 mL autoclave walled with Teflon.Then, 1 g of EA dissolved in 19 mL of water was promptly introduced into the suspension.Ultimately, the solution mixture was sealed and subjected to a temperature of 80 C for a duration of 10 h.After undergoing natural cooling to room temperature, the precipitate was collected and subjected to three thorough washes using ultrapure water and ethanol.It was then frozen at -20 C for 24 h and then freeze-dried at -45 C for 72 h.The resultant composite is denoted as GO-EA in the subsequent sections of the study.

Text S3. Preparation of diethylene triamine (DETA) functionalized graphene oxide (GO)
The GO-DETA composite was synthesized following the technique described by Jiang et al. (Jiang, L. et al., 2023), with some minor adjustments.In summary, a volume of 25 mL of a GO suspension with a concentration of 4 mg/mL was combined with a specific quantity of DETA (8 µL/mL) and agitated.Subsequently, the mixture was transferred to a Teflon reactor capable of withstanding high pressure and subjected to a reaction at a temperature of 90 C for a duration of 16 h, resulting in the production of GO-DETA powder.Subsequently, the GO-DETA was submerged in a 25% (V/V) aqueous ethanol solution for the purpose of eliminating impurities.
Ultimately, the GO-DETA was subjected to a freezing process at a temperature of -20 C for 24 h and then underwent freeze-drying at a temperature of -45 C for 72 h.The resultant composite is denoted as GO-DETA in the subsequent sections of the study.

Text S4. Preparation of hexamethylenetetramine (HMT) functionalized GO
The GO-HMT composite was synthesized following the technique described by Lee et al. (Lee et al., 2012), with some minor adjustments.In summary, 4000 mg of HMT was dissolved in 120 mL of the graphene oxide dispersion, which included about 100 mg of GO.The mixture was stirred magnetically at room temperature for a duration of 10 min.The resultant solution was transported to an autoclave lined with Teflon, which had a capacity of 200 mL.The autoclave was subjected to a constant temperature of 180 C in an electric oven for a duration of 12 h.The autoclave underwent natural cooling following the reaction until it reached the ambient temperature.The black solid obtained was filtered, thoroughly washed with water, and then frozen at a temperature of -20 C for 24 h.Subsequently, freeze-drying was carried out at a temperature of -45 C for 72 h.The resultant composite is denoted as GO-HMT in the subsequent sections of the study.

Text S5. Preparation of Poly(diallyldimethylammonium chloride) (PDDA) functionalized GO
The GO-PDDA composite was synthesized using the approach described by Chen et al. (Chen et al., 2020), with some minor adjustments.In summary, the 1 mg/mL GO dispersion was synthesized as described before.Subsequently, a 5 mL aliquot of PDDA was introduced into the GO dispersion at ambient temperature, and the mixture was agitated for a duration of 4 h.The resultant black solid was obtained using centrifugation, rinsed with water 2-3 times, and then frozen at a temperature of -20 C for 24 h.Subsequently, it was freeze-dried at a temperature of -45 C for a duration of 72 h.The resultant composite is denoted as GO-PDDA in the subsequent sections of the study.

Text S6. Preparation of cetyltrimethylammonium bromide (CTAB) functionalized GO
The GO-CTAB composite was synthesized using the approach described by Yang et al. (Yang and Li, 2014) with minor adjustments.In summary, a solution was prepared by dispersing 15 mg of graphene oxide in 50 mL of water, along with 30 mg of CTAB, and subjecting it to ultrasonic agitation for a duration of 4 h.The suspension was kept undisturbed for a duration of 24 h, during which it naturally separated into distinct layers.Subject the acquired CTAC-GO suspension to centrifugation at a speed of 4500 revolutions per minute for a duration of 15 min.The solid was collected and subjected to freezing at a temperature of -20 C for 24 h, after which it was subjected to freeze-drying at a temperature of -45 C for 72 h.The resultant composite is denoted as GO-CTAC in the subsequent sections of the study.

Text S7. Procedure for performing PFAS adsorption experiment
The adsorption experiments were carried out using 50-mL polypropylene centrifuge tubes (Corning Inc., Corning, NY, USA) in a batch mode.The experiments were repeated three times.
The initial concentration for each PFAS was 10 µg/L, and the dosage of each adsorbent was 100 mg/L.Before introducing an adsorbent, 500 µL of the PFAS solution was extracted from each tube.These samples were used for measurements at a time of 0 min.After the adsorbent was introduced, all tubes were placed on a rotator shaker and set to rotate at a speed of 150 rpm.
Specimens were obtained from each tube at 1, 2, 4, 8, 24, and 48 h.After centrifugation, the supernatant was filtered using 0.2 µm nylon syringe filters, and PFAS were evaluated using an Agilent Technologies 1290 Infinity II LC system combined with a 6470 Triple Quad Mass Spectrometer (LC-MS/MS, Santa Clara, CA, USA).
Additionally, the adsorption of PFAS in river water was examined using a similar technique, except that each PFAS was spiked at 50 ng/L, 200 ng/L, and 1 µg/L.In addition, we assessed the effects of three environmental variables on the adsorption of PFAS.These variables include pH levels ranging from 2 to 12, varying concentrations of natural organic matter (NOM) with humic acid levels ranging from 2 to 100 mg/L, and different ionic strengths with NaCl concentrations ranging from 5 to 200 mM.The initial concentration of PFAS for each type and solution pH were 10 µg/L and 6, respectively.The duration of adsorption was 4 h.The efficacy of the adsorbents in removing PFAS by adsorption was determined by using equation (S1) in the following manner: where Ci and Ct are the PFAS concentration at initial and time (t), respectively.

Text S8. Procedures for characterizing the adsorbents
Zeiss Leo 1550 scanning electron microscope (SEM) equipped with energy dispersive X-ray spectroscopy (EDS) was used to examine the morphology, microstructure, and elemental composition of the absorbents.Functional group analysis in the adsorbent samples before and after PFAS adsorption was carried out using Fourier transform infrared spectroscopy (FTIR; PerkinElmer Spectrum 100, Waltham, MA, USA).Before analysis, the crystal material was cleaned with ethanol and DI water.A certain amount of the sample was placed on the top-plate crystal material of PerkinElmer's Universal ATR (UATR) Accessory.The spectral data were acquired within the 4,000-650 cm -1 spectral band with a resolution of 1 cm -1 .Thermogravimetric analysis (TGA) was conducted using a TGA-5500 instrument (New Castle, DE, USA) in order to assess their thermal behavior.Nitrogen gas adsorption isotherms were obtained at a temperature of 77 K using a 3Flex gas adsorption analyzer (Micromeritics, Norcross, GA, USA).Before the N2 gas adsorption examination, the adsorbents underwent activation under decreased pressure and at a temperature of 50 C (about 0.1 mbar) for 24 h.The Brunauer-Emmett-Teller (BET) technique was used to estimate the specific surface areas, while the pore size distributions were derived using density functional theory.The X-ray diffraction (XRD) was fitted with a graphite monochromator and a D/teX Ultra one-dimensional silicon strip detector.The investigated crystalline samples were pulverized and positioned in zero-background holders, which were subjected to scanning with a 0.01° increment.The Malvern Zetasizer Nano-ZS analyzer (Malvern Panalytical Ltd, Malvern, UK) was used to measure the particle size distribution and zeta potential at a neutral pH and room temperature.PHI Quantera II system was used to perform X-ray Photoelectron Spectroscopy to study the chemical state of the elements present and the bonding configuration at the surface of the absorbent.To include and quantify C along with other elements, a PVA binder was used to drop coat on a Si substrate using the following procedure: 20 mg of sample was added to 100 μL of DI water and ethanol each and sonicated for 15 min.Then, 100 μL of 5% PVA solution was added and sonicated for 15 min.The solution was drop cast on Si substrate and dried for 12 h at 60 C.All the samples were imaged and analyzed at a region where the Si substrate peak was obscure.

Text S9. Chemical analysis
The concentration of PFAS in the river water was measured using EPA Method 537.1.
Concisely, a surrogate of 30 µL (30 ng, 1 mg/L) 13 C-perfluorohexanoic acid (PFHxA) was added to each 400 mL sample.Subsequently, the sample with spikes was introduced into a Hypersep C18 cartridge that had been prepared with methanol and deionized water.Before measurement, the PFAS retained on the C18 cartridges was eluted, and the samples were then enriched with 13 C-PFOS and 13 C-PFOA as internal standards.The Agilent LC-MS/MS was used to quantify the target PFAS in the prepared samples.The supernatant samples obtained from the adsorption experiments were analyzed similarly, except that the solid phase extraction (SPE) step was omitted.Information on PFAS measurements using the LC-MS/MS method may be found in the previous papers (Jiang, T. et al., 2023;Jiang et al., 2022;Zhang et al., 2022a;Zhang and Liang, 2022;Zhang et al., 2022b) as well as in Text S10 and Tables S2 and S3.An analysis was conducted on the anions present in the river water using a 930 Compact IC Flex equipment (Metrohm, Herisau, Switzerland) that was equipped with a conductivity detector.Anions were separated with a Metrosep SUPP 5 column manufactured by Metrohm.The elution technique included using a 1:1 mixture of 1.8 mM Na2CO3 and 1.7 mM NaHCO3 as the eluent, with a 0.7 mL/min flow rate.Later, a solution of sulfuric acid (H2SO4) with a concentration of 0.05 M was used as a regeneration agent to reduce the conductivity.

Text S10. PFAS analysis
Before quantifying PFAS, the samples obtained from the PFAS adsorption studies underwent centrifugation at a speed of 16,000 rpm for 15 min.Following the guidelines of EPA Method 537.1 Revision 2.0 (Zhang and Liang, 2022), the liquid portion of the sample was intentionally contaminated with 13 C4-PFOS and 13 C2-PFOA, which served as internal standards.The target PFASs in the prepared samples were quantified using a 1290 Infinity II LC system paired with a 6470 Triple Quad Mass Spectrometer (LC-MS/MS, Agilent Technologies, Santa Clara, CA, USA).
The experiment used two Agilent Eclipse Plus C18 columns: a ZORBAX analytical column measuring 3 × 50 mm with a particle size of 1.8 μm and a delay column measuring 4.6 × 50 mm with a particle size of 3.5 μm.The columns were operated at a constant temperature of 50 C.The binary mobile phase solvents A and B were selected as ammonium acetate (5 mM) in water and 95% methanol, respectively.The flow rates of these solvents were set at a constant value of 0.5 mL/min.The mobile phase gradient started with a composition of 70% A and 30% B, transitioned to a composition of 0% A and 100% B at 8 min, and maintained this composition for 4 min before returning to the original values.The whole duration was 12 min.
Solubility in water (Sw); Dissociation constant (PKa); N/A: Data not available

Figure S6 .
Figure S6.Zeta potential of the GO and GO-CTAC, respectively.

Table S1 .
Chemicals and reagents used in this study.

Table S2 .
The physicochemical properties of PFAS used in this study.

Table S3 .
Characteristics of the Hudson River water used in this study.

Table S4 .
Dynamic multiple reaction monitoring transitions for the studied PFAS.

Table S5 .
Particle size, BET total specific surface area, and pore size distribution analysis were conducted at 100 °C.

Table S6 .
Summary of binding energy and atomic concentration of GO-CTAC.

Table S7 .
Adsorption performance comparison of the prepared GO-CTAC with other adsorbents reported in the literature.

Table S8 .
Parameters and values derived from adsorption isotherm models of Langmuir, Freundlich, Sips, and Toth for PFAS adsorption by GO-CTAC.

Table S9 .
Experimental and isotherm modeled values of total adsorbed PFAS at equilibrium (qe) and concentrations in aqueous phase at equilibrium (Ce) in the adsorption process by GO-CTAC.