Ternary micro-electrolysis filter media for efficient PFOA removal in water: synthesis, characterization, and performance study Open Access

Polyfluoroalkyl substances (PFAS) are synthetic compounds characterized by hydrogen atoms on a carbon chain being fully replaced by fluorine atoms (O'Connor et al. 2022). They are widely used in consumer and industrial products (Panieri et al. 2022), such as cosmetics, school uniforms, food packaging, and other waterproof materials (Kotthoff et al. 2015; Hoang Nhat Phong et al. 2020; Whitehead et al. 2021; Xia et al. 2022). As persistent organic pollutants, PFAS exhibit bioaccumulation and biomagnification properties, entering the food chain and ultimately affecting human health (Abercrombie et al. 2019; Semerad et al. 2020; Miranda et al. 2021). The US Environmental Protection Agency has set the lifetime health advisory levels for PFOA and PFOS in drinking water at 0.004 and 0.02 ng/L, respectively, (Long et al. 2021; Mukherjee et al. 2023).

PFAS possess hydrophobic carbon-fluorine chains and hydrophilic carboxylic or sulfonic acid functional groups, which result in high solubility and mobility in water (Du et al. 2014; Panieri et al. 2022). Consequently, PFAS, particularly ionic PFOA, are frequently detected in water environments. Current treatment processes struggle to reduce PFAS concentrations to regulatory levels (Leung et al. 2022). PFAS removal methods include biological and non-biological approaches. Biological degradation is still immature, with long degradation cycles. Non-biological methods encompass adsorption, electrochemical oxidation, plasma technology, photocatalysis, and ultrasonic mineralization (Radjenovic & Sedlak 2015; Lin et al. 2016; Aissani et al. 2020; Dibene et al. 2020; Dhore & Murthy 2021; Hu et al. 2022; Kadji et al. 2022; Meegoda et al. 2022). Adsorption is the most common method but requires secondary treatment of the adsorbent. Other chemical degradation methods are costly and mostly experimental (Meegoda et al. 2022). Moreover, PFAS precursors often degrade into other PFAS, with PFOA being a major contributor to environmental detections.

Iron-carbon micro-electrolysis (IC-ME) is an efficient, low-cost technology for treating refractory wastewater, including heavy metals, nitrates, and radionuclides (Li et al. 2021). When materials come into contact with wastewater, micro-galvanic cells form between iron and carbon (Yang et al. 2009; Ruan et al. 2010). The reactive Fe²⁺ and [H] species generated can interact with most pollutants, breaking down organic compounds into less toxic or more degradable substances. ME materials have evolved from early single iron filings to binary and ternary systems. Single ME primarily uses iron shavings or cast iron, with iron as the anode and impurities as the cathode (Feitz et al. 2005). Prolonged use can lead to agglomeration, reducing treatment efficiency. Binary micro-electrolysis, commonly used, involves mixing iron materials with carbon materials (Shen et al. 2019). Nanometer zero-valent iron (nZVI) has replaced conventional iron materials in some studies, offering higher reactivity and removal capabilities. Ternary ME adds a metal catalyst (e.g., Cu, Zn, and Pd) to the binary system, enhancing electron transfer rates and improving efficiency (Xu et al. 2019; Xu et al. 2020; Fu et al. 2021). However, these powdered materials are mainly laboratory-based, difficult to recover, and prone to losing active components, leading to rapid deactivation.

This study employed a liquid-phase reduction method to synthesize highly active nZVI, combining it with high-temperature calcination to prepare granular ternary ME granules (Al/nZVI/C@F) with a particle size of 15–20 mm for PFOA removal. These Al- and nZVI-encapsulated particles prevent leaching into water bodies, reducing ecological risks. Their practical application potential was demonstrated through macroscopic strength tests, microscopic morphology analysis, elemental characterization, and specific surface area measurements. The removal efficiency and application potential of Al/nZVI/C@F for PFOA wastewater were investigated in detail using static batch and dynamic column tests.

• Fly ash was sourced from Shijiazhuang, Hebei Province, China.

• Ferric sulfate heptahydrate (FeSO₄·7H₂O), ammonium carbonate ((NH₄)₂CO₃), and powdered activated carbon, all of the analytical reagent (AR) grade, were procured from Xilong Scientific Co. Ltd.

• Potassium borohydride (KBH₄, AR) and aluminum powder (AR) were obtained from Tianjin Kemiou Chemical Reagent Co. Ltd.

• Montmorillonite powder was acquired from Shanghai Haohong Scientific Co. Ltd.

• Anhydrous ethanol (AR) was purchased from Chengdu Chron Chemicals Co. Ltd.

• Perfluorooctanoic acid (PFOA, sodium salt, GR) was sourced from Shanghai Aladdin Bio-Chem Technology Co. Ltd.

After the reaction, the nZVI + C mixture was filtered and combined with fly ash, montmorillonite powder, and ammonium carbonate to form granules, followed by high-temperature calcination under the N₂ atmosphere. The mass fractions for fly ash, montmorillonite powder, and ammonium carbonate were 50, 10, and 10%, respectively. The calcination was conducted at 800 °C for 1 h, resulting in fly ash-based IC granules (nZVI/C@F). The preparation process for Al/nZVI/C@F granules was similar, with aluminum powder added in an amount equal to nZVI.

Considering that the degradation of PFOA by particles involves active species such as electron generation and HO· radicals, similar to the removal mechanism of methylene blue (MB) (Huang et al. 2022), MB was used as a proxy to quickly evaluate the ME performance of the particles. One gram particles were added to 100 mL MB solution (50 mg/L) and shaken at 25 °C and 180 rpm to test the removal effect of different particles on MB. The concentration of MB was determined by ultraviolet–visible spectrophotometry at 644 nm after sampling at the specified time and filtering with a 0.45 μm filter membrane.

The strength of the granules was measured using a particle strength tester (YHKC-2A). Water absorption was determined according to GB/T 17431.1-2010 and GB/T 17431.2-2010 standards. The microstructure and elemental distribution were observed using scanning electron microscopy with energy-dispersive spectroscopy (SEM–EDS). The composition was analyzed with X-ray diffraction (XRD). Specific surface area and pore size were measured using a fully automated surface area analyzer (Micromeritics APSP 2460) through nitrogen adsorption–desorption at 77 K.

In static batch experiments, 100 mL of PFOA solution at various concentrations were placed in 150 mL conical flasks with granules added at a fixed solid-to-liquid ratio. The flasks were shaken at 180 rpm under different pH and temperature conditions. Samples were taken at predetermined intervals.

The prepared granules were tested as wastewater treatment fillers in a dynamic column setup. The column, made of organic glass with a 26 mm inner diameter and 200 mm length, was equipped with 300-mesh filter membranes at both ends. This setup was used to explore the long-term effectiveness of the granules in removing PFOA.

Samples were filtered through 0.22 μm acetate fiber membranes and analyzed for PFOA using high-performance liquid chromatography (HPLC SPD-16, Shimadzu) under the following conditions: 5 μm C18 column (4.6*150 mm), detection wavelength 210 nm, mobile phase methanol and 20 mM ammonium acetate solution (65:35, v/v), flow rate 1 mL/min, injection volume 20 μL, column temperature 40 °C, and retention time 7.5 min.

Fluoride ions in the degradation process were measured using an ion chromatograph (LC-20ADSP, Shimadzu) with a mobile phase of 1.8 mM sodium carbonate and 1.7 mM sodium bicarbonate at a flow rate of 1.2 mL/min, an analysis time of 10 min, an injection volume of 10 μL, and a temperature of 25 °C. Fluoride ions were detected at a retention time of 3.7 min.

The effect of different fly ash additions on MB removal by nZVI/C@F is shown in Figure 2(b). The results indicate that the fly ash addition had minimal impact on removal efficiency. The introduction of fly ash increased the alumina content, affecting the granule's strength, while also acting as a porogen to enhance porosity. Considering both removal efficiency and mechanical strength, the granules' strength under different fly ash additions is summarized in Figure S1(a). At 20% fly ash addition, the granules had fewer pores and higher strength but reduced contact with pollutants. Increasing fly ash content could reduce the use of the reducing agent potassium borohydride, thus lowering preparation costs. Therefore, 50% fly ash addition was selected as optimal, balancing cost and performance.

The removal efficiency of MB by nZVI/C@F granules prepared at different calcination temperatures is presented in Figure 2(c). Calcination temperature significantly impacted the granule's strength. The strength of ceramic grains under different calcination temperature conditions is shown in Figure S1(b). High-temperature calcination caused the silicon and aluminum oxides to melt, forming a glass phase that enveloped the granule, causing expansion and providing strength and porosity. However, at 1,000 °C, sintering occurred, collapsing pores and reducing the surface area, thereby decreasing MB removal efficiency. No significant difference in removal efficiency was observed at 700, 800, and 900 °C. The particle strength is lowest at 700 °C, making it prone to hardening when applied as a water treatment filler. Therefore, considering energy efficiency and practical application, 800 °C was selected as the optimal calcination temperature.

Figure 2(d) shows the MB removal efficiency of nZVI/C@F granules prepared with different holding times. The best removal efficiencies were observed at holding times of 0.5 and 1 h. As holding time increased, the compressive strength of the material also increased, but its apparent porosity decreased, leading to higher material density. Consequently, increased density reduced the contact between MB and the granules, decreasing removal efficiency. Therefore, a holding time of 1 h was identified as optimal.

In summary, the optimal preparation conditions for nZVI/C@F granules were determined to be an IC ratio of 5:1, a fly ash addition of 50%, a calcination temperature of 800 °C, and a holding time of 1 h.

In this section, the removal effect of PFOA and the preparation cost of granules were compared and analyzed. On the one hand, the granules prepared in this study are a granular filler with a particle size of 15–20 mm, which avoids the problem of easy loss of traditional nZVI when used as a filler, and the filler is easy to recycle. When nZVI/C@F and Al/nZVI/C@F were used as water treatment fillers to treat high-concentration PFOA water, they showed a good removal effect, and no system blockage, filler hardening, or other conventional IC fillers were prone to occur, which could reduce the maintenance and management cost during the operation of the water treatment system. On the other hand, the raw materials required for the preparation of granules in this study mainly include fly ash (0.12 yuan/kg), activated carbon powder (5.2 yuan/kg), potassium borohydride (135 yuan/kg), ferrous sulfate heptahydrate (0.26 yuan/kg), and aluminum powder (about 7 yuan/kg). It is estimated that 61.5 CNY and 31 CNY are needed to prepare 1 kg of nZVI/C@F and Al/nZVI/C@F, respectively, and the commercial IC cost is about 6–7 CNY/kg. Although the preparation cost per unit mass of the two kinds of granules prepared in this study is higher, the power consumption cost will be reduced according to the amount of granules prepared. Therefore, combined with the removal effect of different granules on PFOA, the two granules prepared in this study showed better application value. Since nZVI/C@F has more nZVI and consumes more reducing agents, the cost is higher. After adding Al powder, the cost is reduced, and a better PFOA removal effect is also obtained.

The particle strength and water absorption rates of the prepared nZVI/C@F and Al/nZVI/C@F granules are presented in Table 2. The strengths were 2.58 and 1.81 MPa, respectively, indicating robust granules capable of overcoming the issue of agglomeration common in traditional IC-ME fillers. These strengths meet the requirements for engineering applications such as constructed wetlands. Both types of granules exhibited water absorption rates of around 20%, facilitating thorough contact between the pollutants and the granules, thereby enhancing degradation efficiency.

The Brunauer–Emmet–Teller (BET) analysis results of the particles are summarized in Table 3. The specific surface areas of nZVI/C@F and Al/nZVI/C@F particles were 10.1723 and 48.5195 m²/g, respectively, which were higher than those of fly ash particles prepared from dewatered sludge, clay, and gypsum. The specific surface area of Al/nZVI/C@F is significantly higher than that of nZVI/C@F. This is because the total mass of Al and nZVI in Al/nZVI/C@F is the same as that of nZVI in nZVI/C@F, but the density of Al is much smaller than that of iron. Under the same mass conditions, the volume of aluminum powder is much larger than that of iron, which means Al/nZVI/C@F has more additional powder added, which makes it have a larger specific surface area. This shows that it has good adsorption capacity.

The average pore diameters of 3.1812 ± 0.2125 nm classify the granules as mesoporous materials, with PFOA molecules (approximately 1.1 nm) fitting well within the pores. Thus, while adsorption plays a role in PFOA removal, the primary mechanism is microelectrolysis involving nZVI-C and Al⁰–C, providing pathways for mass transfer and constructing microelectrolysis reaction fields for effective PFOA decomposition.

The static batch experiments investigated the effects of initial PFOA concentration, solid–liquid ratio, initial solution pH, and reaction temperature on the removal efficiency of PFOA by Al/nZVI/C@F granules.

With an initial PFOA concentration of 50 mg/L, a reaction temperature of 25 °C, and a shaking speed of 180 rpm, the solid–liquid ratios were set at 1, 10, 20, and 30 g/L. The removal efficiency of PFOA by the granules is depicted in Figure 5(b). As the amount of granules increased, the removal rate also increased. At a solid–liquid ratio of 1 g/L, the removal rate was only 21.22% due to the limited adsorption and active sites. Increasing the amount of granules increased the collision probability of PFOA molecules with the Fe–C and Al–C interfaces, enhancing the removal rate. When the solid–liquid ratio exceeded 10 g/L, the removal rate surpassed 88.01%.

The initial pH of the solution was adjusted using 0.1 mol/L NaOH or HCl to investigate the effect of pH on PFOA removal. The removal experiments were conducted under the following conditions: an initial PFOA concentration of 50 mg/L, a temperature of 25 °C, and a rotation speed of 180 rpm. The results, presented in Figure 5(c), show that as the pH increased, the removal efficiency decreased. At a pH of 3, the removal rate reached 97.83%, whereas at a pH of 11, the removal rate dropped to 76.48%. This trend can be attributed to the dependance of the ME process on metal corrosion, which varies with pH and influences PFOA removal. Under acidic conditions, the corrosion of Fe⁰ enhances electron transfer, facilitating the formation of numerous ME cells between iron and carbon, thereby promoting the degradation of PFOA. During the anodic reaction, Fe²⁺ is generated, which is subsequently oxidized to Fe³⁺ and precipitates as Fe(OH)_x. Meanwhile, the cathodic reaction produces reactive species such as H· and O·, which further accelerate the degradation of PFOA (Zhang et al. 2018).

The effect of reaction temperature on PFOA removal by Al/nZVI/C@F granules is shown in Figure 7(d), with temperatures set at 15, 25, 35, and 45 °C. Generally, the removal rate decreased with increasing temperature. ME is an exothermic process, so higher temperatures are unfavorable for PFOA removal. Initially, PFOA molecules adsorb onto the granules and then undergo degradation. For a PFOA concentration of 50 mg/L, adsorption is exothermic, favoring lower temperatures. Higher temperatures increase PFOA solubility in water, reducing electrostatic interactions between PFOA and the material. Additionally, higher temperatures increase the vibrational energy of adsorbed PFOA molecules, allowing them to overcome the material's attraction and desorb back into the solution, further reducing the removal rate (Fontecha-Camara et al. 2006; Costa et al. 2012).

During the first 12 days of operation, the PFOA removal efficiency of both columns gradually decreased. This was primarily due to the high initial PFOA concentration (50 mg/L), causing PFOA molecules to continuously enter the columns and first undergo adsorption. This led to a significant number of PFOA molecules adhering to the granule surfaces, occupying active sites, and thus reducing the overall removal efficiency. However, over time, the PFOA molecules adsorbed on the granules began to degrade, freeing up active sites and promoting sustained, efficient, and stable system operation. After 12 days, the removal efficiency stabilized, with the nZVI/C@F column achieving a PFOA removal rate of 64.15 ± 12.09% and the Al/nZVI/C@F column achieving 70.28 ± 8.09%.

The electrode potential of Al (−1.67 V) is lower than that of Fe (−0.44 V). When paired with carbon in a reduction system, the potential difference between aluminum and carbon is greater than that between iron and carbon. This larger potential difference enhances electron transfer, causing aluminum to dissolve more rapidly and facilitating redox reactions. As a result, the Al/nZVI/C@F column exhibits improved processing performance.

Both columns continuously detected fluoride ions in the effluent, indicating ongoing defluorination of PFOA molecules adsorbed on the granules under the influence of Al⁰–C and Fe⁰–C. Over time, the fluoride ion concentration gradually decreased, mainly due to the repeated ME reactions consuming more Fe⁰–C and Al⁰–C, thereby reducing the defluorination rate in subsequent reactions. The fluoride ion concentrations stabilized at approximately 1.66 and 2.87 mg/L in the nZVI/C@F and Al/nZVI/C@F columns, respectively, indicating thorough defluorination of 2.54 and 4.39 mg/L. Given that the actual environmental PFOA concentration is at the ng/L level, applying these materials in real water treatment scenarios would likely result in higher degradation rates and longer operational lifespans.

Experiments were conducted to examine the removal of PFOA using nZVI/C@F and Al/nZVI/C@F granules prepared under optimal conditions. The degradation of PFOA involves the formation of various intermediates, with the ideal outcome being the cleavage of C–F bonds, leading to the release of fluoride ions as the ultimate mineralization product of PFOA (Panda et al. 2019; Schlesinger et al. 2022). Thus, monitoring fluoride ion concentrations during the reaction serves as direct evidence of PFOA degradation.

Two factors likely contribute to the low defluorination rate. First, at high PFOA concentrations (50 mg/L), a substantial number of PFOA molecules adsorbed on the granules' surfaces may hinder the degradation reaction. Second, fluoride ions adsorbed by the granules may reduce the measured fluoride ion concentration. To validate these hypotheses, adsorption and desorption experiments were conducted.

After 480 min of reaction, desorption using methanol revealed PFOA concentrations of 25.4 and 30.2 mg/L for nZVI/C@F and Al/nZVI/C@F, respectively. These results indicate that 50.8 and 60.4% of PFOA were removed via adsorption for nZVI/C@F and Al/nZVI/C@F, respectively. This may be because PFOA molecules are more hydrophobic than hydrophilic (Du et al. 2023), and the use of nZVI/C@F and Al/nZVI/C@F consumes Fe⁰ and Al⁰, producing high-valent cations (Fe³⁺ and Al³⁺) that promote PFOA adsorption (Cai et al. 2022). Combined with the results of fluoride ion determination, it can be calculated that 21.46 and 26.61% of PFOA in n ZVI/C @ F and Al/n ZVI/C @ F were converted into short-chain fluorine-containing intermediates during the removal process, as shown in Figure 9(b).

The high initial concentration of PFOA (50 mg/L) led to significant adsorption on the granule surfaces, hindering the degradation process. Therefore, a lower concentration of 0.5 mg/L PFOA was used in subsequent experiments, with fluoride ion concentrations monitored during the reaction (Figure S3). After 480 min, the fluoride ion concentrations were 0.126 mg/L for nZVI/C@F and 0.147 mg/L for Al/nZVI/C@F, corresponding to defluorination rates of 38.53 and 44.98%, respectively, indicating complete mineralization of 192.7 and 224.9 μg/L of PFOA.

The granules were prepared using fly ash as a base, which has inherent adsorption capacity, especially for fluoride ions. Both types of granules exhibited some degree of fluoride ion adsorption, and due to the larger specific surface area of Al/nZVI/C@F, it showed stronger adsorption at varying initial fluoride ion concentrations. Consequently, the actual defluorination efficiency during PFOA removal by the granules is likely higher than the measured values. Based on the results in Figure S4, the calculated defluorination rates for 0.5 mg/L PFOA are approximately 52.2% for nZVI/C@F and 68.9% for Al/nZVI/C@F. The detailed calculation results of the defluorination rate of 0.5 mg/L PFOA by granules are shown in Table 4.

The above results show that the removal of high-concentration PFOA (50 mg/L) by Al/nZVI/C@F is mainly based on adsorption, and the adsorption of F ions by granules leads to less measured fluoride ions.

The bond energy of the C–F bond is much higher than that of the C–C bond, making it easier to form short-chain fluorocarbon compounds during PFOA degradation. High-performance liquid chromatography-tandem mass spectrometry (HPLC-MS/MS) was used to qualitatively analyze the intermediates in the solution after 480 min of reaction. The total ion chromatogram and corresponding mass spectra (Figure S5) identified the following short-chain perfluorocarboxylic acids: perfluoropropionic acid (C3, m/z = 163), perfluorobutyric acid (C4, m/z = 217), perfluoropentanoic acid (C5, m/z = 263), perfluorohexanoic acid (C6, m/z = 313), and perfluoroheptanoic acid (C7, m/z = 363). Additionally, fluorinated alcohols (e.g., octafluoropentanol, m/z = 231; perfluoropentanol, m/z = 277; pentafluoropropanol, m/z = 147) and alkyl fluorides (e.g., perfluorohexane, m/z = 337) were detected.

In this study, nZVI and Al/nZVI/C@F granules were synthesized using a liquid-phase reduction method combined with high-temperature calcination. The conclusions are as follows:

• The optimum preparation conditions were as follows: Fe:C = 5:1, fly ash addition = 50%, calcination temperature = 800 °C, and calcination time = 1 h.

• The removal efficiency of PFOA by Al/nZVI/C@F (93.90%) was significantly higher than that of commercial IC (12.75%), demonstrating better economic feasibility.

• The highest PFOA removal rate (97.83%) was achieved under the following conditions: initial PFOA concentration = 25 mg/L, solid–liquid ratio = 30 g/L, pH = 3, and reaction temperature = 15 °C.

• After 45 days of dynamic column experiments, PFOA removal rates were 64.15 ± 12.09% for nZVI/C@F and 70.28 ± 8.09% for Al/nZVI/C@F, indicating strong potential for practical applications.

• Al/nZVI/C@F removed high PFOA concentrations (50 mg/L) primarily through adsorption, whereas lower concentrations (0.5 mg/L) allowed for effective defluorination and degradation.

In summary, the ternary ME ceramic Al/nZVI/C@F, prepared in this study, demonstrates significant potential as a filler material for water treatment, addressing PFOA wastewater challenges effectively.

This work is supported by the Guangxi Science and Technology Major Program (AA23073009) and the Innovation Project of Guangxi Graduate Education (YCSW2023294).

All authors contributed to the study's conception and design. Material preparation, data collection, and analysis were performed by S. L., S. Z., J. Z., and Z. W. The first draft of the manuscript was written by L. Z. and S. L. All authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

The data that support the findings of this study are available from the corresponding author upon reasonable request.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.