Graphene Oxide-Constructed 2 nm Pore Anion Exchange Membrane for High Purity Hydrogen Production

Abstract

Alkaline electrolytic water hydrogen generation, a key driver in the growth of hydrogen energy, heavily relies on high-efficiency and high-purity ion exchange membranes. In this study, three-dimensional (3D) wrinkled reduced graphene oxide (WG) nanosheets obtained through a simple thermal reduction process and two-dimensional (2D) graphene oxide act as building blocks, with ethylenediamine as a crosslinking stabilizer, to construct a unique 3D/2D 2 nm-tunneling structure between the GO and WG sheets through via an amide connection at a WG/GO ratio of 1:1. Here, the wrinkled graphene (WG) undergoes a transition from two-dimensional (2D) graphene oxide (GO) into three-dimensional (3D) through the adjustment of surface energy. By increasing the interlayer spacing and the number of ion fluid channels within the membranes, the E-W/G membrane has achieved the rapid passage of hydroxide ions (OH⁻) and simultaneous isolation of produced gas molecules. Moreover, the dense 2 nm nano-tunneling structure in the electrolytic water process enables the E-W/G membrane to attain current densities >99.9% and an extremely low gas crossover rate of hydrogen and oxygen. This result suggests that the as-prepared membrane effectively restricts the unwanted crossover of gases between the anode and cathode compartments, leading to improved efficiency and reduced gas leakage during electrolysis. By enhancing the purity of the hydrogen production industry and facilitating the energy transition, our strategy holds great potential for realizing the widespread utilization of hydrogen energy.

1. Introduction

Under the threat of energy shortage and environmental degradation, the development of sustainable and clean energy is increasingly urgent [1]. Hydrogen-based clean energy has attracted more and more attention because of its advantages such as large reserves, low pollution, high efficiency, high specific energy, and sustainable development [2,3,4]. Importantly, the rising demand for hydrogen energy necessitates the continuous improvement of hydrogen production technology systems, focusing on enhancing efficiency, reducing costs, and ensuring high-quality hydrogen production [4,5]. Nowadays, hydrogen production by electrolysis of water has become a highly promising technology because water is a renewable raw material [6]. Alkaline anion exchange membranes (AEMs) have attracted significant attention for water electrolysis due to their ability to produce high-purity hydrogen (up to 99.9% vol) without relying on noble metal catalysts [7,8]. However, the commercialization of AEM-based electrolyzers still faces several technical challenges. These include the need for membranes with high hydroxide ion conductivity, excellent mechanical and chemical stability under strongly alkaline conditions, low gas crossover rates, scalable fabrication, and cost-effective raw materials. Therefore, developing low-cost and high-performance AEMs is key to advancing industrial hydrogen production. Graphene oxide (GO)-based membranes have shown promise due to their rich oxygen-containing functional groups and lamellar structure, which are conducive to hydroxide ion transport and tunable nanochannel formation. Nevertheless, existing GO membranes face critical drawbacks. The spontaneous stacking of GO nanosheets via van der Waals interactions leads to narrow interlayer spacing, which limits ion conductivity and reduces effective surface area. Furthermore, the structural instability of GO in alkaline media can deteriorate long-term performance. These limitations highlight the need for novel structural designs and functionalization strategies to overcome these barriers and meet the rigorous requirements for commercial AEM applications. Alkaline anion exchange membranes (AEMs) water electrolysis is expected to become a mainstream hydrogen production strategy due not only to its low-cost production but also high purity of the generated hydrogen (up to 99.9% vol) [9,10]. Hence, designing and developing AEMs with excellent hydroxide conductivity, isolation of hydrogen and oxygen from cathodes and anodes, and stability in alkaline aqueous solutions is very important [11,12,13,14].

Commercially available AEM, such as Fumasep FAA-3-50 [15,16,17], are designed to exhibit specific characteristics that balance the isolation of hydrogen and oxygen gases while facilitating the passage of OH⁻. However, the high cost of commercial AEMs like FAA-3-50 can contribute to the overall cost of hydrogen production through electrolytic water [18]. To achieve low-cost hydrogen production, researchers have explored various alternative membrane materials and structures [19]. Among them, the good hydrophilicity of 2D GO sheets can provide a carrier for the transport of OH⁻ and enhance ion conduction [20]. In addition, its unique nanosheet morphology can precisely regulate the pore size through layer-by-layer (LBL) assembly. Hence, graphene oxide (GO)-based membrane has emerged as an excellent candidate material in the field of AEMs [21,22,23,24]. For example, Zhan et al. reported an anion exchange membrane with sulfonated reduced graphene oxide modification layers, which improve monovalent anion selectivity and controllable resistance [25]. However, in the process of anion exchange membrane preparation, the irreversible accumulation of 2D nanosheets caused by Van der Waals forces [26], which greatly reduces the effective surface area, hinders the transport of electrolyte ions, and affects the performance of the anion exchange membrane.

In order to obtain complete synergistic properties in the macroscopic volume, it is often necessary to integrate two-dimensional nanosheets into three-dimensional (3D) porous nanosheet networks, also known as three-dimensional monolithic structures (such as 3D foams and 3D aerogels) [26,27]. The 3D overall structure can prevent surface-to-surface accumulation of two-dimensional nanosheet structures, and its interconnection network can not only provide channels for electron transport, but also provide continuous cavities or channels for electrolyte ions diffusion [28,29]. Guan et al. fabricated a 3D novel membrane with 2D channels by embedding nanoporous crystals into reduced GO membranes. The method greatly improved fluid permeability [26].

Therefore, in this work, 3D wrinkled graphene as a spacer and graphene oxide were connected by amide bonds and LBL assembled on polyethersulfone (PES) membrane (named E-W/G membrane) by a simple pressure filtration method (Figure 1). Noteworthy, the above strategy has the following advantages: (1) 3D wrinkled graphene can effectively increase the spacing between graphene oxide layers, precisely control the nanopore size to 2 nm, which is beneficial for reducing the gas crossover, H₂ and O₂, across the E-W/G membrane. (2) Ethylenediamine as a crosslinking agent to connect 3D wrinkled graphene and graphene oxide through amide bonds, which creates water permeation channels, and improves ion transport capacity. (3) Amide reaction can also enhance the stability and durability of the prepared membrane. Therefore, the E-W/G membrane offers several advantages, including efficient gas isolation, high current density, and low preparation cost.

2. Experimental Section

2.1. Materials and Chemicals

Commercial PES membrane (pore size: 0.22 μm) was purchased from Zhongli Filtration Equipment Co., Ltd. (Haining, China). Commercially available FAA-3-50 membrane (thickness 50 μm) was purchased from Fumasep Co., Ltd. (Bietigheim-Bissingen, Germany). GO (oxygen content: ≤40.0%, average lateral size: <4 μm, monolayer probability: <90%) was purchased from Sixth Element Materials Technology Co., Ltd. (Changzhou, China). Potassium hydroxide (KOH, >95%), manganous sulfate (MnSO₂, >99%), and hydrogen peroxide (H₂O₂, AR) were purchased from Kelong Chemical Co., Ltd. (Chengdu, China), and ethylenediamine monohydrate was purchased, respectively, from Shanghai Titan Science Co., Ltd. (Shanghai, China).

2.2. Preparation of WG

The decrease in oxygen content increases the surface energy of graphene oxide, causing the crystal lattice to wrinkle [30,31]. The slurry containing graphene oxide flakes (0.05 mg mL⁻¹) was lyophilized into the form of aerogel, warmed up for 24 min (20 °C per minute), and then thermally reduced in an air atmosphere for 2 min. WG, which has undergone heat treatment to minimize oxidation (0.1 g), was dispersed in a solution of 33.33 mL of deionized (DI) water, 1.89 g of H₂O_2, and 0.011 mg of MnCl₂ for two days to introduce carboxyl groups at the lattice edges. It was centrifuged and freeze-dried after that.

2.3. Preparation of E-W/G Membrane

Ethylenediamine was added to 4 mL of GO solution to form a mixture (final concentration 0.4 mg/mL), diluted to 50 mL with DI water, and stirred for 2 h to make ethylenediamine adsorbed on the oxygen-containing functional groups [32]. To investigate the optimal component ratio for structural stability and ion transport, WG solutions were mixed with GO at volume ratios of 1:1 and 2:1, respectively. The 1:1 ratio enables uniform dispersion and stable amide bonding formation, while the 2:1 condition enriches the 3D wrinkled phase, enhancing porosity and reducing gas crossover. This comparative strategy provides insights into the structure–property relationship of the membrane system. Mix thoroughly, and add DI water to 200 mL. Then, 15 min of ultrasonication followed by filtration using PES nanofiltration membranes with a specialized extraction apparatus must be achieved. The membrane is then heated for 2 h in an air environment at 80 °C in a blast drying oven.

2.4. Characterization of E-W/G Membrane

On the Elementar, Vario EL Cube, X-ray Photoelectron Spectroscopy (XPS) (Thermo Fisher Scientific, Waltham, MA, USA) experiments (K-Alpha, Al Kα excitation source) were made for chemical compositions. A Perkin Elmer Frontier mid-IR FTIR spectrometer was used to record the FTIR (Fourier transform infrared spectroscopy) spectra (PerkinElmer, Inc., Waltham, MA, USA). Using a Rigaku MiniFlex 600, wide-angle X-ray diffractometry was carried out across a range of 5–85°. Membrane surface and cross-sectional morphologies were characterized using scanning electron microscopy (SEM, FEI Inspect F50, PANalytical B.V; Almere, The Netherlands). Thermo Fisher’s Dxr2xi spectroscopy was used to obtain the Raman spectra, which had an excitation wavelength of 532 nm (Thermo Fisher Scientific; Waltham, MA, USA). The contact angle was measured on an optical tensiometer contact angle measuring instrument (Bilin Scientific) (Beijing Hake Test Instrument Factory; Beijing, China). The D-spacing of membranes was studied by the X-ray diffractometer (XRD, PANalytical, Almere, The Netherlands, Cu Kα radiation source). Brunauer–Emmett–Teller surface area analysis (BET) was carried out using a Kubo (X1000) (Beijing Beiaode Electronic Technology Co., Ltd.; Beijing, China).

2.5. Properties Test of E-W/G Membrane Electrolytic Water and Gas Cross-Testing

The H-type electrolytic cell configuration with the specified electrode materials and membrane allows for efficient water electrolysis. The working electrode (NF) facilitates the electrochemical splitting of water into hydrogen and oxygen gas, while the counter electrode (graphite rod) supports the corresponding reduction and oxidation reactions. The reference electrode (Hg/HgO) provides a reference potential for accurate measurement and monitoring of the electrochemical process. For the water electrolysis test, an electrochemical workstation (CHI-660E, Shanghai CHI Instruments Co., Ltd., Shanghai, China) was employed. The resultant gas is driven via a desiccant (to eliminate the impact of water vapor) and onto a gas chromatograph (GC2060, Shanghai Keen Instrument Co., Ltd., Shanghai, China), which examines the gas composition using a gas thermal conductivity detector (TCD) attached to the gas chromatograph utilizing argon (Ar, 5 mL min⁻¹) as a carrier gas (Tables S2–S6 shows the accuracy and accuracy of gas chromatography standard gas test). In the experimental analysis, the results are presented as signal peaks generated by fitting the obtained data. These signal peaks correspond to specific gases or components present in the samples. To determine the gas composition and ratios of the products, the peak areas of standard gases with known volume ratios are compared with the peak areas of the fitted products.

3. Results and Discussion

Heating the single molecular layer of graphene oxide enables the formation of a 3D wrinkled structure, known as WG, while the stacking of the graphene oxide molecular layer and the 3D wrinkle layer offers precise control over the membrane’s pore size. This precise control allows the membrane to efficiently isolate gases and prevent any crossover between them. The introduction of the amide group results in a positive charge within the internal transmission channel. This positively charged environment facilitates the accelerated transit of OH⁻ ions, ensuring a high current density. Additionally, it strengthens the bonding force between the layers, further enhancing the overall performance of the system. Figure 2a–c exhibit the surface morphology of the GO membrane, WG membrane, and E-W/G composite membrane, respectively. It is evident that after cross-linking, the membranes exhibit a flatter state. Furthermore, it provides visual evidence of the successful introduction of amide groups into the membranes. Extended stirring ensures the stability of the amide group, enabling uniform bonding with the carbon atoms. As a result, the nitrogen element is evenly distributed throughout the nano-porous E-W/G membrane shown in Figure 2d. In the FT-IR spectra of WG/GO and E-W/G, as depicted in Figure 2e, noticeable absorption peaks at 1621 cm⁻¹ can be observed, indicating the presence of the amide group. These findings suggest that C-N covalent bonds are formed through the reaction between the amide monomer and the oxygen-containing functional group present in GO [33]. The reaction involves nucleophilic addition reactions between the amide and epoxides, as well as condensation reactions of the amide with hydroxyl groups. As a result of the involvement of GO’s hydroxyl group in the condensation reaction, the absorption peak of E-W/G nearly disappears, indicating successful bond formation [34,35].

X-ray photoelectron spectroscopy (XPS) was employed to analyze the chemical composition and bonding characteristics, as depicted in Figure 2f. The C1s portion of the XPS spectra was deconvoluted to provide a more detailed elemental analysis of the composite membranes. It was observed that E-W/G exhibit similar peaks at 288, 286, and 284 eV, respectively [36,37], corresponding to C=O, C-O, and C-C [38,39,40,41]. A new absorption peak appears near the energy level of 285 eV, where the C-O bond drops significantly and C-N increases compared to the C1s spectrum of the WG/GO membrane in Figure S1b. These results are associated with the nucleophilic substitution reaction that causes the amine to attack the three-membered epoxide ring of GO; consequently, the epoxide ring loop was opened, forming covalent bonds between C-O and C-N [42].

In WG, the residual defects resulting from thermal reduction lead to a conversion of marginal sp³ carbon and certain oxygen-containing groups into carboxyl groups [36,43]. This conversion results in an increase in the C=O content within the material. When ethylenediamine interacts with the oxygen-containing functional group of GO, condensation and nucleophilic reactions occur between the amine and the epoxy, leading to the formation of a C-N covalent bond. As a result, the epoxide ring is opened, and covalent bonds are formed between C-O and C-N. This reaction causes a slight decrease in the peak corresponding to C-O and the appearance of a new peak corresponding to C-N in the spectrum. Table S1 presents the elemental content as determined by XPS analysis, providing a clearer understanding of the elemental changes. Additionally, the pore size distribution in the range of 1.8–2.1 nm was primarily determined by BET, as shown in Figure 2g. The existing literature suggests that pores with a size of 2 nm are highly suitable for the transport of OH⁻ [44]. The choice of a ~2 nm pore size is based on the critical balance between ionic conductivity and gas barrier performance. On one hand, pores below 2 nm effectively suppress the diffusion of diatomic gas molecules such as H₂ (kinetic diameter ≈ 0.289 nm) and O₂ (≈0.346 nm), thereby minimizing gas crossover. On the other hand, such pore sizes are still sufficiently large to allow the transport of solvated hydroxide ions (OH⁻), which possess a hydration shell diameter of ~0.3–0.6 nm. This selectivity ensures efficient ionic conduction while maintaining excellent gas separation. Furthermore, the dense and tortuous nanochannel structure formed by stacked GO/WG layers provides additional physical resistance and electrostatic repulsion to gas molecules, contributing to the outstanding gas barrier performance of the E-W/G membrane. The pore size of the membrane used in fuel cells is usually greater than 4 nm, and a smaller pore size is more conducive to a gas barrier [45].

The closely laminated nano-channels comprising GO and WG exhibit efficient blocking of gas passage in the electrolytic water reaction, while allowing unobstructed passage of hydroxide ions. This is possible because the amide group, which is tightly bound to the layer structure and carries a positive charge, facilitates the rapid flow of OH⁻. By controlling the charge on the GO surface or between layers, a strong electrostatic attraction can be generated, enhancing the penetration and transfer of OH⁻, introducing amide groups, introducing more functional sites on the graphene oxide sheet, and enhancing the permeability of OH⁻. The analysis of d-spacing based on XRD results, as shown in Figure 2h, reveals that the introduction of WG in comparison to GO leads to an increase in d-spacing from 9.2 Å to 9.5 Å. Furthermore, the E-W/G samples exhibit a larger d-spacing of 11.77 Å. This stretching of the layer spacing can be attributed to the breaking of π-π and hydrogen bonds during the formation of C-N covalent bonds. The contact angle of the E-W/G composite membrane measures 55.41°, which is similar to that of the commercial FAA-3-50 membrane shown in Figure 2i. EDA monomer involved in covalent and reduction reaction reached an equilibrium state, protonated amines and hydrophilic oxygen-containing functional groups showed better hydrophilic performance than the GO membrane. The enhanced hydrophobicity of the E-W/G membrane, compared to the pure GO membrane, can be attributed not only to the presence of the polar carboxyl group in WG but also to the introduction of the amide group.

To evaluate the purity of electrolyzed water in engineering applications and measure the gas crossover rate, hydrogen permeability is monitored at the anode, while oxygen permeability is detected at the cathode. The permeability value, defined as the gas crossover rate, is determined by analyzing and fitting the signal peaks. E-W/G membranes were employed as separators in H-type cells [46]. The testing was conducted at a temperature of 30 °C, using a 1 M KOH solution, and a current density of 10 mA/cm⁻² [47]. The aim was to determine the extent of gas crossover and ensure the purity of the generated gases. The presence of a large number of -COOH groups in the E-W/G membrane creates a hydrophilic environment, as shown in Figure 4f. Under hydrophilic conditions, at a constant current density is 100 mA cm⁻², the purity of hydrogen produced by electrolysis of E-W/G membrane was tested, and the data of continuous electrolysis were recorded in Figure 3. Here, the hydrogen purity test was carried out in five rounds, each cycle lasting 12 h. We used the composite membrane E-W/G, W/G, and FAA-3-50 as separators in an H-type cell to verify the possibility of directly producing high-purity hydrogen by electrolysis of water. Hydrogen purity is tested once after each continuous electrolytic cycle. The purity of hydrogen and oxygen was tested; after signal peak fitting, the purity of hydrogen was >99.9%. In contrast, the hydrogen purity fitted using the commercial FAA-3-50 membrane is slightly greater than 99.4%, but does not reach 99.9%. In some applications, the purity level of 99.4% does not meet the purity standards for direct use like fuel cells. The permeability resistance of the E-W/G membrane with 2 nm channel is better than that of the FAA membrane, and the crossover rate of the E-W/G membrane shows a trend of decreasing with the extension of time, because the GO layer is compressed under the action of current, and the interlayer structure becomes more compact, which also weakens the gas penetration. The solution enters between the GO and WG layers, causing the hydrogen bonds and π-π bonds to break. As a result, the layer spacing is compressed to just enough to allow the passage of OH⁻ ions while effectively blocking the crossover of gases. This mechanism allows E-W/G membranes to have a gas barrier efficiency of >99.9% compared to an efficiency of ~99.4% for commercial FAA-3-50.

Figure 4a,b exhibit SEM images depicting the evolution of the E-W/G composite membrane throughout prolonged electrolysis conducted under continuous constant current conditions. These images vividly illustrate that the membrane’s surface undergoes a transformation, becoming smoother and more uniform. Additionally, the previously observed layer structure becomes denser and more tightly packed as the electrolysis process continues, signifying a noticeable enhancement in the membrane’s overall stability. Our findings reveal a noteworthy reduction in the overall roughness of the sample. Our AFM analyses in Figure 4c,g–i. Figure 4i shows the values of the root square roughness (Rq), the average roughness (Ra), and the mean roughness depth (Rz) of E-W/G before and after 48 h CCE. Indicate that the initial surface roughness in the entire scan area measures 254 nm. However, after subjecting the membrane to 48 h of electrolysis, the scan area surface roughness undergoes a discernible transformation, decreasing to 156 nm. After 48 h of electrolysis, XPS studies show a decrease in C-O bond content, indicating GO reduction during electrolysis. This is probably because increasing hydrophilicity has changed the surface shape. The presence of polar carboxyl groups and -COOH, -NH functions on the membrane’s surface guarantees its hydrophilicity, as do decreased C-O bonds, decreased layer spacing, increased roughness, and decreased contact angle individually. Notably, a slight shift in the C-N peak position is observed between Figure 2f and Figure 4d. This difference can be attributed to changes in the chemical environment of nitrogen-containing functional groups after prolonged electrolysis. Specifically, the electrochemical treatment may lead to partial protonation or rearrangement of amide groups, altering the local electron density and thus causing binding energy variation in the XPS spectra. Additionally, hydration effects and structural compaction after 48 h of electrolysis could also influence the surface potential and electronic interactions within the membrane matrix. In Figure 4e, this experiment provides evidence that the membrane retains its hydrophilic nature even after undergoing continuous electrolysis, which can be attributed to the persistent presence of polar groups on its surface [48,49]. After 24 h of continuous electrolysis, the contact angle decreased to 49.46°, and after 48 h, the decrease became smaller, to 44.48°. This is because the protonated amine grafted on the GO surface is more hydrophilic than the hydrophilic oxygen-containing group on the GO surface during long-term electrolysis. Raman spectroscopy, as depicted in Figure 4f, can also provide evidence of the aforementioned phenomenon. Carboxyl groups enhance its dispersibility in water; the combination of hydrophilicity, electrical conductivity, and structural stability makes the E-W/G membrane promising for various applications.

4. Conclusions

We have successfully developed an innovative E-W/G anion-exchange membrane by integrating WG sheets into graphene oxide sheets. By precisely controlling the nanopore dimensions to 2 nm, facilitating unimpeded movement of OH⁻ ions, and introducing positively charged amide groups within the layers, the E-W/G membrane demonstrates an improvement over the commercial FAA-3-50 membrane in reducing crossover rates of H₂ and O₂ gases, enhancing purity in water electrolysis applications. These attributes make the E-W/G membrane a promising candidate for various electrolysis applications. Furthermore, the E-W/G membrane shows remarkable stability and high current density in extended testing, holding potential for direct hydrogen applications and making significant strides in advancing sustainable energy. This breakthrough not only propels membrane technology but also lays the foundation for a greener and more energy-abundant future.