Fluorine-Doped Graphene Oxide-Modified Graphite Felt Cathode for Hydrogen Peroxide Generation

Abstract

Electrochemical oxygen reduction via the two-electron pathway (2e-ORR) is an emerging method for producing H₂O₂. It is cleaner, safer, and more convenient compared to the anthraquinone process. Graphite felt is one of the cathode candidates for large-scale cells due to its excellent mechanical properties. However, commercial graphite felt often fails to achieve the desired hydrogen-peroxide yield because of its low catalytic selectivity for the 2e-ORR pathway. Fluorine-doped carbon materials are expected to enhance 2e-ORR selectivity. This is because the electronic structure of carbon atoms adjacent to fluorine atoms may facilitate the production of hydrogen peroxide while hindering its further reduction. In this study, fluorine-doped graphene oxide (FGO) was prepared by the hydrothermal method. Subsequently, graphite felt modified with FGO was fabricated and used as the cathode for H₂O₂ production. The results indicated that in alkaline media, the graphite felt modified with FGO achieved a catalytic selectivity of 93% and a generation rate of 8.91 mg cm⁻² h⁻¹. In comparison, commercial graphite felt had a catalytic selectivity of 75% and a generation rate of 2.10 mg cm⁻² h⁻¹. Moreover, graphite felt modified by FGO also exhibited excellent electrocatalytic performance for H₂O₂ generation in neutral media. This research provides a fundamental study to promote the application of graphite felt in the environmentally friendly electrocatalytic production of hydrogen peroxide in industries.

1. Introduction

The electrochemical synthesis of H₂O₂ through the two-electron oxygen reduction reaction (2e-ORR) pathway is an emerging hydrogen-peroxide production method, which can be operated under mild reaction conditions, offering greater cleanliness and safety [1,2]. Moreover, it serves as an effective means to achieve on-site production for the utilization of H₂O₂ solution [3,4,5]. However, the production efficiency of the electrochemical method is still much lower than that of the state-of-the-art anthraquinone process.

To enhance the production efficiency of H₂O₂ in a large-scale cell requires attempting to obtain a more concentrated H₂O₂ solution in a shorter time and with less electricity consumption. In other words, it means improving the production rate and current efficiency. The key to improving current efficiency lies in the selectivity of the catalyst or electrode material. Hence, it is crucial to design and develop active electrocatalysts with selectivity for H₂O₂. The 2e-ORR pathway for the reaction of O₂ molecules on the cathode competes with the thermodynamically favorable 4e-ORR pathway to H₂O. OOH* is a common intermediate of both the 2e and 4e pathways. The magnitude of the adsorption energy of OOH* determines the selectivity of H₂O₂ generation [6]. Therefore, electrocatalysts with appropriate adsorption-free energies of OOH* are expected to improve H₂O₂ selectivity.

In recent years, carbon materials have attracted much attention as the cathode material for H₂O₂ production because of their advantages, such as low cost, high storage capacity, and modifiability [7,8,9,10]. Carbon materials, including carbon nanotubes, graphene, mesoporous carbon, activated carbon, and others, have been doped with oxygen functional groups [11,12,13] and heteroatoms, such as N [14,15,16,17], B [18,19,20], S [21,22,23], and others. In these materials, heteroatoms combine with the carbon skeleton, resulting in charge redistribution in the carbon material and modulating the adsorption-free energy of oxygen and the reaction-intermediate OOH* during the reaction process. As a result, the activity of the ORR and the selectivity of H₂O₂ are improved, and so is the yield of H₂O₂ in the electrolysis cell.

Moreover, in order to enhance the electrochemical generation of hydrogen peroxide, it is crucial to increase the current density. This requires the use of appropriate substrate materials to facilitate the attainment of the desired current density. Graphite felt (GF) has been widely used as an electrode material due to its large specific surface area, excellent mechanical properties, and good electrical conductivity [24,25]. These characteristics make it an ideal substrate material for cathode electrodes that can withstand high current densities in large-scale applications. Currently, graphite felt is being researched for its potential use as a cathode for the direct production of hydrogen peroxide. However, its selectivity for 2e-ORR is limited, leading to lower production yields [26]. Therefore, modifying graphite felt with high-performance catalysts and using it as the cathode could be one of the approaches to enhance the electrochemical production efficiency of H₂O₂.

Fluorine atoms doped in carbon materials can enhance the O₂ adsorption energy to improve the ORR electrocatalytic activity of carbon materials [27,28,29]. It is found that F doping elongates the O=O bond distance and shortens the C-O distance, thus clearly demonstrating its superiority in O₂ adsorption on fluorine-doped carbon materials [30]. In addition, the doped F species with different configurations have a different influence on the selectivity for the ORR pathway. Based on density functional theory (DFT) studies and experimental results, Zhao et al. discovered that the doped F species with different configurations have a different influence on the selectivity for the ORR pathway. That is, the covalent –CFx species (–CF₂ and –CF₃) can promote H₂O₂ synthesis by selectively catalyzing the two-electron pathway of oxygen reduction to promote the synthesis of H₂O₂ [31]. Furthermore, Wang et al. observed that different kinds of F–CNTs had different electrocatalytic properties, which might be related to the contents of –CF₂ and –CF₃ [32].

In this study, fluorine-doped graphene oxide (FGO) was synthesized as an electrocatalyst by introducing –CF₂ groups to graphene oxide (GO) via the hydrothermal method, using GO as the substrate and hydrofluoric acid as the F-source. The synthesized FGO was then used for the preparation of FGO-modified graphite felt cathode material, which exhibited high hydrogen-peroxide selectivity and effective yields in alkaline or neutral media.

2. Results and Discussion

2.1. Characterization of FGO and FGO/GF

Many methods were used to identify the existence and the chemical speciation of F in FGO. In the X-ray powder diffraction (XRD) pattern of GO (Figure 1a), the weak broad peak centered at approximately 2θ = 19.8° originates from the carbon skeleton structure. The intensive diffraction peak at 2θ = 9.8° is related to the intercalation of a large number of oxygen functional groups [33], such as –OH, –COOH, and C–O–C. It was found that the peak at 9.8° disappeared and a broad peak at about 24° was observed in the XRD pattern of FGO, indicating that most of the oxygen functional groups in GO had been removed after the hydrothermal treatment with hydrofluoric acid. The peak at about 2θ = 18° was observed in FGO, which may be related to the reduction of oxygen functional groups and the doping of F [34].

The Fourier transform infrared (FTIR) spectra also confirm the introduction of F in FGO compared with that of GO (Figure 1b). The characteristic bands at 1627 cm⁻¹ and 1725 cm⁻¹ in the FTIR spectra of GO are, respectively, assigned to the stretching vibration of C=C and C=O, which originated from the graphene-oxide-like benzene ring structure [35]. The broad bands appeared in the region of 3700–2400 cm⁻¹, and the absorption peak at 1040 cm⁻¹ contributed to O–H stretching and vibration of C–O. For FGO, the characteristic peaks related to the oxygen functional groups weakened or disappeared, and a new peak at 1100 cm⁻¹ could be assigned to the C–F bond connected to a nearby sp² hybridized carbon atom [36]. Furthermore, the presence of F atoms and their uniform distribution in FGO is also evidenced in EDS mapping images (Figure 1c). The nitrogen adsorption–desorption isotherms of GO and FGO are shown in Figure 1d. The BET surface area of GO and FGO was calculated to be 204.6 m² g⁻¹ and 237.4 m² g⁻¹, respectively. The isotherms for GO and FGO exhibited typical type IV curves with type H3 hysteresis loop [37], indicating the existence of micro- and meso-pores. From the inset pore-size distribution in Figure 1d, it can be seen that the pore size of FGO is obviously smaller than that of GO. This may be attributed to the strong noncovalent interactions formed during the hydrothermal treatment of GO sheets under acidic pH conditions. Such interactions lead to the aggregation of FGO, resulting in the formation of many small holes with thick walls [38].

The chemical speciation of F in FGO was further confirmed by X-ray photoelectron spectroscopy (XPS). In the XPS spectrum of FGO (Figure 2a), the F1s peak appeared at 685 eV, and the XPS data showed the presence of carbon, oxygen, and fluorine with atomic percentages of 88.3%, 7.77%, and 3.93%, respectively. This indicates that the F element has been successfully introduced into graphene oxide after hydrothermal treatment with HF. As shown in Figure 2b, the C1s spectra of GO can be fitted with four peaks, centered at about 284.4, 284.8, 286.7, and 288.3 eV, corresponding to the C=C, C–C, C–O, and –C=O groups, respectively [39]. After the HF treatment, the relative intensity of C=C and C–O decreased, and a new peak appeared at 291 eV, which is sure to be related to the C–F bond.

In general, there are three chemical forms of C–F bonds that may be formed in F-doped carbon materials; these are ionic C–F, semi-ionic C–F, and covalent C–F [40,41]. Figure 2c,d show the high-resolution XPS spectra of F 1s and C 1s for FGO, respectively. In the high-resolution F 1s spectra, the prominent peak at approximately 688.6 eV is associated with the –CF₂ bond [42,43,44], while the peaks around 685.5 eV and 691.3 eV correspond to the semi-ionic C–F bond and the covalent –CF₃ bond, respectively [45,46,47,48]. Additionally, the broad peak centered at 291 eV in the high-resolution C 1s spectrum provides further evidence that –CF₂ is the primary bonding configuration of F [49,50].

Based on the above results, it can be inferred that most of the F atoms in FGO exist in –CF₂ bonds. The fluorine atoms have strong electronegativity, which attracts the surrounding electrons and leads to a change in the charge distribution of the neighboring carbon atoms on the graphene-oxide surface.

To clarify the morphological characteristics of the FGO modified GF electrodes, the SEM and EDS mapping images of GF and FGO/GF were analyzed, as shown in Figure 3. The pristine GF surface is smooth and almost defect-free with an interconnected network structure under different magnifications (Figure 3a). In contrast, the GF modified by FGO impregnation (Figure 3b) shows that the dispersed catalyst was locally agglomerated and adhered to the fiber surface of the GF. The results of EDS indicated that after the modification of graphite mats with FGO, the contents of C, O, and F on the surface of the graphite-mat fibers were 89.3 at%, 7.4 at%, and 3.3 at%, respectively (Figure 3c). Due to the large specific surface area provided by the dense nature of the graphite-mat fibers, the catalyst is able to attach to as many fibers as possible to expose the active sites.

2.2. Electrochemical Characterizations

Figure 4a,d show the RRDE polarization curves in O₂-saturated 0.1 M KOH solution or 0.1 M Na₂SO₄ solution with the FGO or GO loaded on the disk electrode and the rotating rate of electrode of 1600 rpm. As shown in Figure 4a, the onset potential (E_onset, at H₂O₂ current density at 0.1 mA cm⁻²) of 0.72 V for FGO in 0.1 M KOH was more positive than that of GO (0.59 V), indicating that the ORR catalytic activity of FGO is higher than that of GO. In addition, FGO exhibited a significantly better H₂O₂ selectivity than GO in the potential range of 0.2–0.7 V (~80–90% for FGO, ~70–80% for GO), in which the transferred electrons number was calculated as 2.13 at 0.70 V. Similarly, the hydrogen-peroxide generation performance of FGO in neutral electrolytes (0.1 M Na₂SO₄, pH ~ 7) was also significantly increased compared to that of GO, with the measured onset potential positively shifted by about 100 mV (Figure 4d) and the selectivity increased from ~60% for GO to ~80% for FGO (Figure 4f).

These observations suggest that F-atom doping in graphene oxide facilitates the enhancement of electrocatalytic activity and hydrogen-peroxide selectivity for the O₂ reduction reaction, which may be attributed to the fact that the –CF₂ doping into the carbon skeleton modulates the adsorption-free energies of graphene oxide for O₂ and the intermediate OOH* [31].

2.3. Electrogeneration of H₂O₂

The electrochemical hydrogen-peroxide-production performance of FGO/GF was evaluated by using it as cathode in an H-type cell with 1.0 M KOH solution or 0.1 M Na₂SO₄ solution, and the concentration of H₂O₂ in the electrolyte after 1200 s of electrolysis was used as the evaluation index. The results in the alkali electrolyte showed that the Faraday efficiency (FE) of FGO/GF was 93% at 0.7 V (Figure 5a), which is significantly higher than that of the commercial GF and GO/GF. Furthermore, the current density during the electrolysis using FGO/GF as a cathode was significantly enhanced to 20.3 mA cm⁻² at 0.3 V (Figure 5b). Finally, 371 mg L⁻¹ of H₂O₂ was yielded in 1.0 M KOH with a Faraday efficiency of 70% and current density of 20.3 mA cm⁻² (Figure 5b,c). In the neutral electrolyte, FGO/GF also exhibited good performance: the FE reached ~77% at 0.2 V, and the current density reached ~16 mA cm⁻² at 0 V. As a result, 165 mg L⁻¹ of H₂O₂ was generated using FGO/GF as a cathode, while those levels were 1.28 mg L⁻¹ and 93 mg L⁻¹ using commercial GF and GO/GF as a cathode (Figure 5d–f).

The results of constant potential electrolysis demonstrated that fluorine-doped graphene oxide significantly enhanced the electrocatalytic production of hydrogen peroxide by graphite felt. This improvement was attributed to the excellent oxygen reduction reaction (ORR) activity and two-electron selectivity of FGO.

The H₂O₂ yield was also evaluated with different FGO loadings (1.2 and 4 mg cm⁻²) on GF in 1.0 M KOH solution (Figure 6a–c) and 0.1 M Na₂SO₄ solution (Figure 6d,e). The graphite felts with 1 mg cm⁻² FGO loading had the highest hydrogen-peroxide yield in 1.0 M KOH, and the highest H₂O₂ production rate was calculated to be 8.91 mg·h⁻¹·cm⁻². In 0.1 M Na₂SO₄, the graphite felt with 2 mg cm⁻² FGO loading had the highest hydrogen-peroxide yield of 257 mg L⁻¹ at 0.1 V, and the maximum production rate was calculated to be 6.18 mg·h⁻¹·cm⁻². This may be due to the fact that the ORR process is limited by the mass-transfer process, where the diffusion of O₂ may be relatively slow or the adsorption of OOH* on the catalyst surface may be weak. An appropriate increase in catalyst loading can increase the contact opportunity between O₂ and the catalyst and improve the reaction rate.

Compared to the electrocatalytic GF cathodes recently reported in the literature, FGO/GF has a significant advantage in terms of H₂O₂-generating capacity (

Table 1. Comparison of H₂O₂ electro-generation with other studies.

Cathode	pH	Current Density (A m−2)	t (min)	[H2O2] (mg·h−1·cm−2)	Refs
Graphite felt	3	132	300	0.11	[51]
GF-(4)	7	10	120	0.74	[52]
Modified graphite felt	7	50	60	2.5	[53]
gc-GF	9	67	60	6.89	[54]
FGO/GF	7	203	20	6.18	This work
FGO/GF	13.7	166	20	8.91	This work

), providing greater current density and yield of hydrogen peroxide.

The stability test was conducted by repeating the constant potential electrolysis test under the optimized conditions in the alkali electrolyte (Figure 7). By replacing the electrolyte and cleaning the electrodes every 20 min at 0.3 V vs. RHE, pH = 13.7, the current density and yield of H₂O₂ were tested. During the five cycles, the current density was relatively stable, and the yield of hydrogen peroxide decreased slightly but remained within a reasonable range, with the yield of hydrogen peroxide during the last cycle maintained at 91%.

3. Materials and Methods

3.1. Reagents and Chemicals

Graphene oxide (GO) was purchased from Suzhou TanFeng (Suzhou TanFeng Graphene Technology Co., Ltd., Jiangsu, China), graphite felt (GF, G475) was purchased from SCI Material Hub, potassium hydroxide (KOH, AR) and sodium sulfate (Na₂SO₄, AR) were purchased from Shanghai Aladdin (Shanghai Aladdin Bio-chemistry Technology Co., Ltd., Shanghai, China), and aqueous hydrofluoric acid (HF, 40 Ltd., Beijing, China) and aqueous hydrofluoric acid (HF, 40 wt%) were purchased from Beijing Bailing Wei (Beijing Bailing Wei Technology Co., Ltd., Beijing, China).

3.2. Material Synthesis

3.2.1. Synthesis of FGO

A total of 35 mg of GO and 1 mL of hydrofluoric acid (40%) were added to 35 mL of deionized water and ultrasonicated for 20 min until homogeneous. A total of 30 mL of the mixture was transferred to a 50 mL PTFE-lined reactor and heated to 150 °C for 24 h. After natural cooling to room temperature, the product was filtered through a microporous membrane, washed with distilled water, and dried to obtain FGO.

3.2.2. Preparation of Cathodes

The graphite felts (GFs) were first cut into several 1 cm×2 cm strips, washed with ethanol and deionized water by ultrasonication for 30 min, and then dried at 60 °C for 24 h. The catalyst inks were obtained by adding 1 mg/2 mg/4 mg of GO or FGO to 1 mL/2 mL/4 mL of deionized water and 5 wt% Nafion solution (volume ratio of 0.995:0.005) and ultrasonicated for 1 h. The inks were drop-coated onto the GF fibers, and then dried at room temperature to obtain GO/GF or FGO/GF.

3.3. Characterization of Catalysts and the Composite Cathodes

The X-ray powder diffraction (XRD) patterns were acquired by the Bruker AXS D8 Advance instrument (40 kV, 40 mA), using Cu Kα (λ = 1.5406 A) radiation (Bruker, Billerica, MA, USA). Fourier transform infrared (FTIR) spectra were recorded on a Nicolet iS20 spectrometer (Thermo Nicolet, Waltham, MA, USA). The scanning electron microscopy images were obtained using a field-emission scanning electron microscope (Zeiss supra 55). The N₂ adsorption–desorption isotherms were studied on a Micromeritics (Micromeritics, ASAP 2020, Waltham, MA, USA) analyzer to calculate specific surface area and porosities of graphene-based materials using the multi-point Brunauer–Emmett–Teller (BET) technique. X-ray photoelectron spectroscopy (XPS) data were obtained using a Thermo Scientific K-Alpha instrument with an Al Ka X-ray source, and the peak areas were fitted using the software XPSPEAK 4.1 by Lorentz (30%) and Gaussian (https://gaussian.com/) (70%) functions until a fitting criterion of c2 less than 10 was achieved.

3.4. Electrochemical Performance Tests

The typical electrochemical study was performed on a CHI760E electrochemical workstation (Chenhua, Shanghai, China) in a standard three-electrode system. For GO and FGO catalysts, the catalytic activity of the catalysts was evaluated using a theoretical collection efficiency of 37% rotating ring-disc electrode (RRDE). A modified glassy carbon electrode (0.2475 cm²) loaded with a GO or FGO catalyst was used as the working electrode, the counter electrode was a platinum sheet (1 cm × 1 cm), and the reference electrode was a saturated calomel electrode (SCE). The catalyst ink was prepared by dispersing 1 mg GO or FGO in 1 mL deionized water and 5 μL 5 wt% Nafion solution. A total of 8 μL of catalyst dispersion was applied dropwise to the glassy carbon electrode at 700 rpm and dried. The electrochemical tests were carried out at ambient temperature and a pressure of (0.1 M KOH) at pH ~ 13, and (0.1 M Na₂SO₄) at pH ~ 7, respectively. The potentials mentioned in this study have been converted for comparison with the Reversible Hydrogen Electrode (RHE) using Equation (1), which is as follows:

E_RHE = E _SCE + 0.0592 × pH + 0.2438

(1)

For the rotating ring-disc electrode (RRDE), the average number of transferred electrons (n) and the H₂O₂ selectivity were calculated by the following equation:

$$n={\displaystyle \frac{4\times I_d}{I_d+I_r/N}}$$

(2)

$${\mathrm{H}}_2{\mathrm{O}}_{2 }(\%)={\displaystyle \frac{200\times I_r/N}{I_d+I_r/N}}$$

(3)

where I_r is the ring electrode current, I_d is the disc electrode current, and N is the ring electrode collection efficiency (0.37). The ring current was recorded by keeping the Pt ring at 1.2 V (vs. RHE) throughout the test.

3.5. Electrogeneration of H₂O₂ and Analytic Methods

The H₂O₂ yields of GO- and FGO-modified graphite felt cathodes in (1 M KOH) at pH ~ 13.7, and (0.1 M Na₂SO₄) at pH ~ 7, respectively, were determined by constant potential electrolysis in an H-type electrolysis bath with a Nafion diaphragm, in which a catalyst-loaded graphite felt was used as the working electrode, a Pt plate as the counter electrode, and an SCE connected to the cathode chamber by a salt bridge as the reference electrode. The cathode chamber volume was 8 m L for both. The electrolysis was conducted in a constant potential model with O₂ gas bubbling near the cathode for 20 min. The generated H₂O₂ in the electrolyte after electrolysis was determined by the Ce(SO₄)₂ method, which, using a UV Vis spectrophotometer (UV1100, Shanghai, China), measured at 400 nm. The current efficiency (FE) for H₂O₂ production was calculated as the following equation [55]:

$$\mathrm{F}\mathrm{E}(\%)={\displaystyle \frac{n\mathrm{F}\mathit{cV}}{\int Idt}}\times 100\%$$

(4)

where n is the number of electrons transferred by the reduction of oxygen to H₂O₂, F is Faraday’s constant (96,485 C mol⁻¹), c is the concentration of H₂O₂ (M), V is the volume (L), I is the current (A), and t is the electrolysis time (s).

The stability test was conducted by repeating the constant potential electrolysis test under the same conditions. The electrolyte was replaced and we cleaned the electrodes every 20 min.

4. Conclusions

In this study, GO was treated with HF under hydrothermal conditions to obtain F-doped GO, and the main chemical speciation of F elements in FGO was confirmed as –CF₂ groups. The introduction of –CF₂ groups into GO led to a significant improvement in its catalytic selectivity for 2e-ORR. Then, the FGO-modified graphite felt (FGO/GF) was fabricated and its performance in producing H₂O₂ was evaluated by using it as the cathode in alkaline and neutral electrolytes. The results showed that the H₂O₂ yield and current efficiency of the FGO/GF cathode were significantly improved compared with those of commercial GF and GO/GF in both alkaline and neutral electrolytes. A total of 371 mg L⁻¹ of H₂O₂ was generated after electrolysis at 0.3 V vs. RHE in 1.0 M KOH for 1200s with a current density of 20.3 mA cm⁻². The present work provides a facile strategy for the preparation of efficient, low-cost, heteroatom co-doped cathode materials for the electrocatalytic production of hydrogen peroxide.