The electro-Fenton process (E-Fenton) [1-4], one of the advanced oxidation processes (AOPs), has been shown to be a promising technology for degrading refractory pollutants in wastewater [5-10]. The process is a simple electrochemical reaction process. First, H₂O₂ is generated from the reduction of O₂through a 2e^- process without additional acquisition, shipment, or storage (Eq. 1). Subsequently, Fe²⁺ is added to the system to react with the H₂O₂ to generate hydroxyl radicals (·OH) (Eq. 2). The hydroxyl radicals, which have a high electrode potential (2.8 V vs. standard hydrogen electrode (SHE)), can rapidly non-selectively destroy refractory pollutants[11-13].

(1)

(2)

In the E-Fenton process, H₂O₂ is first generated at the cathode through the 2e^- oxygen reduction reaction (ORR) process [14, 15]. So, improvement of the efficiency of ORR for H₂O₂ production is crucial for E-Fenton systems [16]. The cathode material can greatly influence the mechanism and kinetics of ORR [17, 18]. At present, carbonaceous materials have been widely investigated for their excellent ORR reactivity, with advantages such as high overpotential for hydrogen evolution, non-toxicity, low cost, and good stability. Carbonaceous materials such as graphite, carbon felt, graphite felt, carbon sponge, activated carbon fiber, and carbon-polytetrafluoroethylene (PTFE) composite electrodes are widely used in E-Fenton systems [19-25]. Among them, graphite felt (GF), already a commercialized cathode material, shows promising advantages such as a 3D interconnected structure that provides abundant active sites and rapid mass-transfer, outstanding anti-corrosion resistance, mechanical integrity in flexible electrodes, and ease of manufacture for large-scale applications [21]. Accordingly, GF has attracted great attention in the electrocatalysis field. Because the kinetics of the electrochemical process of ORR on carbon is complicated and quite sensitive to surface properties, various pretreatment methods have been introduced to make carbonaceous materials more efficient for H₂O₂ production and the E-Fenton process. As reported, the electrocatalytic activity of GF can be significantly enhanced through surface modification [26].

Recently, it has been widely reported that N-doped carbon materials possess high ORR activity, attracting much attention for ORR electrocatalysts for fuel cell applications. Among them, electrocatalysts synthesized using non-precious metals and polymer precursors containing nitrogen atoms have shown promising activity and unique durability for ORR [27-29]. Polyaniline (PANI), which contains a favorable combination of aromatic rings connected via nitrogen-containing groups, is widely used as both an N and C precursor. For example, Wu et al. [30] reported electrocatalysts with high ORR performance derived from iron, cobalt, and PANI. Kim et al.[31] reported NH₃-activated Fe/PANI composites for ORR. However, until now no study has been devoted to the performance of PANI derived material in an E-Fenton system.

In the present investigation, we focused on the fabrication and evaluation of a polyaniline coated graphite felt (PANI@GF) composite for the electrocatalytic reduction of oxygen and subsequently as an E-Fenton cathode. The PANI@GF composite was synthesized through a novel electro-polymerization method. Both the surface chemistry and structural properties of the composite were studied, and its electrocatalytic activity for ORR was investigated. Herein, dimethyl phthalate (DMP) was used as a model substrate to evaluate the performance of PANI@GF for the E-Fenton process.

DMP, a representative di-alkyl phthalate ester, is widely used as an indispensable additive for plastics to increase their flexibility. The annual world production of plastics is estimated to be around 100 million tons. As a result, a great amount of DMP is drained into the aquatic environment through the disposal of manufacturing wastewater. DMP belongs to the endocrine-disrupting chemicals, whose ubiquity in the environment has brought great concern to the public and environmental researchers [32, 33]. DMP has been listed as a priority pollutant in many countries and is used in the rapid performance assessment of some AOPs, including the E-Fenton reaction.

All chemicals used in this work were of analytical grade. The commercial GF (Hunan Jiuhua Carbon High-Tech Co. Ltd., China) was first degreased with acetone in an ultrasonic bath for 15 min and then washed with deionized water several times to remove the residual acetone. To synthesize PANI@GF, a GF plate (2.0 cm × 2.0 cm) was used as the substrate, while a Pt foil (2.0 cm × 1.0 cm) and a saturated calomel electrode (SCE) were used as the anode and the reference electrode, respectively. A PANI layer was electro-polymerized onto the surface of the GF in an aqueous solution containing 0.2 mol/L aniline and 0.75 mol/L H₂SO₄ by the galvanostatic method at 0.5 mA/cm² for 1 h at room temperature. Next, the plate with PANI layer was subjected to heat treatment at 900℃ for 3 h under N₂ atmosphere. The heat-treated product was then washed with 0.1 mol/L aqueous H₂SO₄ several times to remove any unstable and ORR-nonreactive residues. The as-prepared PANI@GF composite was used as the cathode of the E-Fenton system in the following experiments.

The morphology of the PANI@GF was observed using scanning electron microscopy (SEM; Quanta 400F, FEI/Oxford/HKL, Holland and France). X-ray photoelectron spectroscopy (XPS) characterization was performed on an ESCALAB 250 (ThermoFisher Scientific, USA).

The DMP degradation experiment was carried out in a three-necked cell at a cathodic potential of 0.5 V (vs. SCE) using PANI@GF or GF as the E-Fenton cathode, Pt foil as the counter electrode, and a SCE as the reference electrode. A 50 mg/L DMP solution was prepared in advance. FeSO₄·7H₂O solution was added to the DMP solution and the pH was adjusted to the required value using dilute sulfuric acid and sodium hydroxide. The reaction started when H₂O₂ was electrochemically generated in-situ. After a certain reaction time, 1.00 mL of the reaction solution was removed from the system and placed in a centrifuge tube containing 0.02 mL methanol for later analysis.

The H₂O₂ concentration of the reaction solution was determined using the potassium titanium oxalate method with a UV-VIS spectrophotometer (TU1810, Universal Analysis, Beijing, China). The DMP concentration was measured using a high performance liquid chromatograph (HPLC; LC 2140, Techcomp, Shanghai, China) equipped with a reverse phase column (XT erra MS C-18, 5 μm, Waters, USA) and a UV detector. The mobile phase was a mixed solution of 50% acetonitrile and 50% water (V/V), and the detection wavelength was set to 276 nm. The total organic carbon (TOC) concentration was determined with a TOC/TNb analyzer (Vario TOC cube, Elemental, Germany) using the standard non-purgeable organic carbon (NPOC) method.

The morphology and composition of the electro-polymerized GF were investigated by SEM and XPS. Fig. 1(a) and (b) show that the untreated graphite felt had a 3D porous structure stemming from its overlapped carbon fibers. The surface of the carbon fibers was very smooth with occasional small attached lumps. These attachments may have been some impurities and could be eliminated through pretreatment with acid and alkali. Fig. 1(c) and (d) show that after electro-polymerization, polyaniline was well deposited on the carbon fibers, making their surface rough. The covering itself was loose and had a porous structure, as demonstrated in Fig. 1(e) and (f), which was expected to help enhance the electrochemically active area and be beneficial for the diffusion of O₂, thereby enhancing the ORR mass-transfer process and performance and then the E-Fenton performance.

The atomic configuration and stoichiometric composition of PANI@GF were analyzed by XPS. The atomic content of the PANI@GF was calculated from the integrated peak area of the C1s, O1s, and N1s signals divided by their atomic sensitivity factor. The XPS survey scan revealed that the N and O content of the composite was 1.96 at.% and 7.91 at.%, respectively. The atomic O content of PANI@GF was decreased compared with that of the untreated GF (13.57 at.%), indicating the depletion of oxygen functional groups, as demonstrated in Fig. 2(a). The N1s spectra depicted in Fig. 2(b) show that the PANI@GF composite contained four different kinds of N coordinations: pyridinic N, amine N, pyrrolic N, and graphitic N. The content of pyridinic N and pyrrolic N in PANI@GF was obviously higher than the others. These results confirm that, as well as the macroscopic 3D porous structure of GF, the material also had a micro three-dimensional porous structure, and that N atoms were successfully doped into the GF surface.

As shown in Fig. 3, it took 2 h for GF to completely degrade DMP, but only 45 min for PANI@GF to completely degrade DMP. This indicates that the DMP degradation performance of the E-Fenton system could be significantly enhanced using the PANI@GF cathode. According to some reports [34], the atomic O content of graphite felt increases after treatment by chemical or electrochemical methods. It is generally believed that an increase in oxygen containing functional groups is the main reason for the improvement of modified GF. In this work, the atomic O content of the PANI@GF electrode was lower than that of the GF electrode, suggesting that it was not the oxygen functional groups but the macroscopic and microscopic porous structure and the N atom doping that enhanced the DMP degradation performance of the E-Fenton system with PANI@GF cathode.

The degradation conditions had a significant influence on the performance of the PANI@GF cathode, so the optimum conditions for DMP degradation were further studied. The carbonization temperature is important for the structure of the PANI@GF surface. Within a certain temperature range, the DMP degradation ability of PANI@GF was gradually improved as the carbonization temperature was increased (Fig. 4). PANI@GF treated at 900 ℃ had the strongest DMP degradation ability. This enhancement might be attributed to the higher carbonization temperature causing the polyaniline on the surface of GF to form a looser and more porous structure. This would have improved the O₂ mass transfer process and thus the DMP degradation performance of the cathode.

Just like the carbonization temperature, the polymerization time also affected the surface structure of PANI@GF. As shown in Fig. 5, the DMP degradation performance of GF alone was not good. After 1 h reaction 40% DMP still remained. When the electro-polymerization time was 1 h, it took only 45 min to completely degrade DMP at the PANI@GF cathode. When the electro-polymerization time was extended to 2 h, however, the ability of the cathode to degrade DMP declined. This behavior can be explained as follows: when the polymerization time is too short, polyaniline cannot adhere well to GF. In contrast, when the polymerization is too long, the resulting thick polyaniline layer could seal the micro pores on the surface of the GF and consequently hinder the diffusion of O₂, which would cause the DMP degradation performance of the composite to deteriorate.

The DMP degradation conditions, including O₂ flow rate, Fe concentration and pH value, were also studied. The DMP degradation performance of the composite under different O₂ flow rates is shown in Fig. 6. When no O₂ purge of the reaction cell was carried out, only 20% of the DMP was degraded in 1 h. When the O₂ flow rate was increased to 0.2 L/min, the DMP degradation performance was significantly improved. This is because when no O₂ was fed into the cathode, the concentration of O₂ dissolved in the solution was low, and the ORR was limited. When the O₂ flow rate was increased, the DMP degradation performance was accordingly improved. However, the DMP degradation rate observed at O₂ flow rate of 0.6 L/min was slower than that at 0.4 L/min. As the O₂ flow rate was increased, the ORR rate-limiting step changed from O₂ concentration to O₂ diffusion. The concentration of O₂ in the solution reached its maximum value when the O₂ flow rate was 0.4 L/min. Subsequently, increasing the flow rate beyond 0.4 L/min could not further increase the O₂ concentration. On the contrary, bubbles generated by the larger flow rate scoured the PANI surface, resulting in a decline in DMP degradation performance. Hence, the 0.4 L/min O₂ flow rate was the optimal one.

The influence of Fe²⁺ concentration on DMP degradation is shown in Fig. 7. Varying the Fe (Ⅱ) concentration had little influence on the DMP degradation results at pH=3.0. This might be attributable to the fact that Fe (Ⅱ) can be recycled through a variety of paths in the E-Fenton system. Furthermore, Fe (Ⅲ) can also be reduced to Fe (Ⅱ) at the cathode. As a result, 1.0 mmol/L Fe (Ⅱ) was sufficient for DMP degradation in this system.

The pH plays a fundamental role in the E-Fenton system. As depicted in Fig. 8, a pH of 3.0 was optimal for the present system, consistent with the 2.0-3.0 range for the E-Fenton system and that of conventional Fenton systems [35]. Under neutral or near neutral conditions, the DMP degradation performance was very poor, which might stem from a lack of Fe²⁺[36]. The pH of the system can affect the chemical form of Fe. When the pH is greater than 4.0, Fe exists in the form of hydroxides such as [Fe (OH)]²⁺, [Fe (OH)₂]⁺, and [Fe₂(OH)₂]⁴⁺. This leads to the precipitation of ferric hydroxides, hindering the conversion to Fe²⁺ and resulting in poor performance. Additionally, as a weak acid, H₂O₂ decomposes easily at higher pH, which is inimical to DMP degradation.

A novel three-dimensional porous composite PANI@GF electrode was successfully prepared and employed for DMP degradation in aqueous solution. The PANI@GF electrode had both a macro and micro three-dimensional porous structure and was successfully doped with nitrogen atoms, which are beneficial for O₂ diffusion and ORR to electrochemically generate H₂O₂. The apparent rate constant for DMP degradation at the PANI@GF cathode was 0.0753 min^-1, five times higher than that for the unmodified graphite felt cathode. It is believed that the macroscopic and microscopic porous structure and the N doping of PANI@GF enhanced its ORR performance. The optimal carbonization temperature and electro-polymerization time were 900℃ and 1 h, respectively. Both factors affected the structure of the PANI@GF electrode, and thus its DMP degradation performance. Higher carbonization temperature facilitated the formation of a looser microscopic pore structure that likely improved O₂ diffusion. The pH of the system was found to be of crucial importance for DMP degradation, and the optimal value was 3.0. A suitable O₂ flow rate was beneficial to DMP performance, and was found to be optimal at 0.4 L/min. Because Fe²⁺ can be recycled through a variety of paths in the E-Fenton system, a concentration of 1.0 mmol/L was adequate. PANI@GF is an attractive alternative cathode for E-Fenton systems applied to the degradation of organic contaminants in wastewater.