High performance of membrane capacitive deionization with ZnS/g-C₃N₄ composite electrodes

With the development of industry and the growth of population, the global demand for freshwater is rapidly increasing, and the original scarce freshwater resources are more tense. Since seawater and brackish water account for more than 96% of total water resources on the earth (Shiklomanov 2000), obtaining fresh water from seawater or micro-salt water has attracted considerable attention, which is supposed to be a feasible strategy to relieve the crisis of fresh water shortage. Traditional desalination technologies, such as reverse osmosis, multi-stage flash distillation, and electrodialysis processes have showed the disadvantages of excessive energy consumption, high cost, and low efficiency, which hinders their more widespread application (Mohamed et al. 2020). New desalination techniques that are energy efficient and cost-effective are explored all the time.

Capacitive deionization (CDI) technology has been a research hotspot in the area of seawater or brackish water desalination in recent years (Chung et al. 2020; Torkamanzadeh et al. 2020; Gong et al. 2021), which shows the advantages of low energy consumption, ecological friendliness, high separation, simple design, and easy operation (Yan et al. 2018). The technique is simple and easy to operate. Desalting can be carried out at low pressure and room temperature, and the applied voltage at both ends of the electrode generally does not exceed 1.6 V. Unlike reverse osmosis or distillation-based desalination systems, CDI does not need to be coupled to a high-pressure pump or heat source, allowing easy system scaling (Suss et al. 2015). Additionally, CDI is environmentally favorable because energy, rather than chemicals, is used to power the operation process. Therefore, it does not cause secondary pollution to water bodies, and is a clean and efficient desalination technology. This is mostly a physical process that allows CDI equipment to have a long service life and low maintenance costs (Porada et al. 2013). During the desalination process, saline ions are attracted and stored in the electrical double layer of charged electrodes, and released from the surface of discharged electrodes. Electrode material is crucial for the performance of CDI as its structure, porosity, surface and electrochemical properties greatly determine the ions adsorption efficiency (Peng et al. 2011; Dlugolecki & van der Wal 2013). Carbon materials, such as activated carbon, mesoporous carbon, carbon nanotubes, carbon aerogels, and graphene are mainly used for CDI electrode preparation because of the large surface area and wide pore size distribution (Huang et al. 2017; Leong et al. 2019). However, drawbacks of low conductivity, irregular pore shape and size distribution of these carbon-based electrodes limit the CDI desalination capacity (Sufiani et al. 2019). Therefore, exploring new electrode materials for high performance of the CDI system are necessary.

The graphite phase carbon nitride (g-C₃N₄) is a novel non-metallic semiconductor material with the merit of environment friendly large specific surface area, good electrical conductivity, etc. (Zheng et al. 2012). g-C₃N₄ has caused broad interest in the scientific community as photocatalyst. Also, it is begun to be used as electrode active materials in recent years due to its good conductivity and hydrophilicity. For example, Wang et al. (2018) used g-C₃N₄ as electrode materials for CDI, showing an excellent salt adsorption capacity and cyclic stability. Kavil et al. (2018) prepared nano-g-C₃N₄/MnO₂ and g-C₃N₄/SnO₂ composites for supercapacitor applications, and the g-C₃N₄ improved the specific surface area and promoted the electron transport at the electrode-electrolyte interface. Thiagarajan et al. (2020) prepared NiMoO₄/g-C₃N₄ for supercapacitors, which showed a large specific capacitance (501 F/g) and high stability, maintaining a capacity of up to 91.8% even after 2,000 charge and discharge cycles. In a word, g-C₃N₄ as electrode materials could improve the conductivity, hydrophilicity, and ion transmission of the electrodes.

Metal sulfides (MSs) are a kind of potential electrode materials in the metal matrix. MSs tend to cause sulfur defects in the formation process, thus forming a large number of active sites, which facilitate the catalytic reaction (Shiraz et al. 2021). Zinc sulfide (ZnS) has attracted extensive attention due to its advantages such as high capacity and density, rich sulfur and zinc resources, low cost and less pollution, and has become a cathode material for high performance of energy storage (Guo et al. 2018; Xie et al. 2018). In addition, ZnS is an important semiconductor material with wide energy gap (3.5–3.8 eV) and high specific capacitance (926.3 mAh/g) (Du et al. 2017; Xu et al. 2018). However, disadvantages of ZnS, such as poor conductivity and volume expansion have been found as well during its application. Recombination with other materials is supposed to be a way to make up for the deficiency. For example, Sarma et al. (2015) synthesized a flower-shaped ZnS/TiO₂ nanotubes composite electrode, showing high specific capacitance and stability under large charging conditions.

In this study, g-C₃N₄ was used as the carrier to support ZnS nanoparticles, and ZnS/g-C₃N₄ composite was prepared by a high-temperature calcination method. The morphology and structure of the composite was analyzed, and the electrode was prepared. The electrochemical characteristics of the composite and the desalination performance were studied. The results show that ZnS can form a spherical particle structure on g-C₃N₄ lamellar layer, which weakens the stacking effect between the carbon nitrification lamellar layer and improves the capacitance of the material, thus improving the capacitive deionization performance of the electrode. The power supply voltage, inflow velocity and initial concentration of NaCl solution all have an effect on the desalting capacity of the capacitor. Through different CDI and MCDI mode tests, ZnS/g-C₃N₄//ZnS/g-C₃N₄ (MCDI) maximum adsorption capacity is 27.65 mg/g.

Thiourea was chosen as the experimental precursor. To ensure a well-mixed product, 5 g of thiourea was weighed in a mortar and ground for 30 min. After that, it was placed in a quartz boat and then into a tubular furnace (SK-G08123K, Tianjin Zhonghuan Electric Furnace Co., Ltd, China). It was heated to 550 °C at 10 °C/min in a nitrogen atmosphere for 2 h. The light yellow g-C₃N₄ powder was successfully prepared after being cooled to room temperature.

The precursor of choice was thiourea. To create a homogenous mixture, 1 g of zinc acetate and 5 g of thiourea were weighed and mortared for 30 min. The quartz boat of the tubular furnace should then be used. It was heated to 550 °C under nitrogen environment at a rate of 10 °C/min and maintained for 2 h. To obtain orange ZnS/g-C₃N₄ powder, the mixture was cooled to room temperature.

First, ZnS/g-C₃N₄ complex, acetylene black, and polyvinylidene fluoride were ground evenly in a mortar with a mass ratio of 8:1:1. Acetylene black was used as conductive agent and polyvinylidene fluoride was used as adhesive. Then drop N-methyl pyrrolidone as a solvent, stirring for more than half an hour, to obtain an uniform slurry. Using graphite paper as fluid collector, the electrode paste was coated on graphite paper by a simple coating method, and the electrode was obtained by drying it in a drying oven at 60 °C for 12 h.

The microstructure and morphology of the samples were investigated by scanning electron microscopy (SEM, Zeiss MERLIN Compact) and transmission electron microscopy (TEM, FEITecnai G2 F30). X-ray diffractometer (XRD, Rigaku Ultima IV) was used to analyze the crystal structure of the material. Fourier transform infrared spectrometer (FTIR) was used to analyze the structure of the materials, mainly used to determine the surface functional groups of the samples. Nitrogen adsorption and desorption isotherms were measured at 77 K. Before the measurement, the samples were degassed at 463 K for 6 h in vacuum using the McAsap2460 instrument. The Brunauer–Emmett–Teller (BET) method was used to calculate the specific surface area and pore volume. A Barrett–Joyner–Halenda (BJH) model was used to derive the pore size distribution from the adsorption branch of the isotherm. The water contact angle was measured using JY-82C Kruss DSA on the powder sample after compression. Titration of ultrapure water was done automatically using equipment containing 1.6 μL of water droplets.

In this experiment, electrochemical workstation (CHI660E, Chenhua, China) was used to test the electrochemical performance of samples with a three-electrode system. In 1-M NaCl aqueous solution, a three-electrode system was used at room temperature. g-C₃N₄ and ZnS/g-C₃N₄ composites were used as working electrodes. Before the electrochemical performance test, the working electrodes were immersed in NaCl for 12 h to fully soak the electrolyte solution. Platinum (Pt) electrode and saturated calomel electrode (SCE) were used as the counter electrode and reference electrode, respectively.

The CDI cell operates in flow mode, which means that two electrodes are parallel to each other with a gap between them, allowing brine to flow between them. The NaCl solution was directly driven up and down by a peristaltic pump (BT100-2J, Chinese long pump). The solution entered the device through the lower port and exited through the upper port. The electric field was powered by a constant voltage DC power supply (VICTOR 3003A, China). A conductivity meter was used to continuously record the change in conductivity of the NaCl solution (DDSJ-308F, INESA Scientific Instrument Co. Ltd, China). The concentration of NaCl solution was calculated using the measured conductivity and a standard conductivity-concentration curve.

In this experiment, NaCl solutions of various concentrations were prepared using ultrapure water with conductivity values below 3 μS/cm. At a steady temperature, the conductivity and salt solution concentration have a positive relationship. Supplementary material, Figure S1 depicts the connection between conductivity and the concentration of NaCl solution. By using linear fitting, it was possible to derive the function relationship between solution concentration and conductivity value as y = 1.9375x + 19.226 with correlation coefficient R² = 0.9997, which is a satisfactory fit. It was discovered that, in the solution concentration range of 0–2,000 mg/L, the concentration of NaCl solution showed a good linear association with the conductivity. The results of the solution's measured conductivity can be used to determine how much desalination is required given changes in the solution's concentration.

The morphology of g-C₃N₄ and ZnS/g-C₃N₄ composites was characterized by SEM and TEM. Supplementary material, Figure S2(a) and (b) shows that the pure g-C₃N₄ has a porous and layered stack structure with smooth surfaces. From the SEM images of ZnS/g-C₃N₄ composites (Supplementary material, Figure S2(c) and (d)), it can be observed that spherical-like nanoparticles of ZnS were uniformly distributed on the g-C₃N₄ sheet, and this was further verified by TEM images (Supplementary material, Figure S2(e) and (f)). The g-C₃N₄ phase facilitates the ZnS particles precipitate and restrain their agglomeration, which contributes to a high specific surface area of ZnS. Besides, the porous structure of the composites could be beneficial to the ions diffusion during the desalination process.

Supplementary material (Figure S3) shows the XRD patterns of pure g-C₃N₄ and ZnS/g-C₃N₄ composites. For g-C₃N₄, a weak peak at 13.1° and a strong peak at 27.4° are observed, which correspond to (100) and (002) planes of g-C₃N₄ crystal, respectively (JCPDS No. 87-1526) (Liao et al. 2012). For ZnS/g-C₃N₄ composites, there are four additional diffraction peaks at approximately 28.5°, 33.1°, 47.5°, and 56.3°, corresponding to (111), (200), (220), and (311) planes of the hexagonal crystal of ZnS (JCPDS No. 05-0566). The observed diffraction peaks of both g-C₃N₄ and ZnS confirmed the successful synthesis of the ZnS/g-C₃N₄ composite.

The FTIR spectra of g-C₃N₄ and ZnS/g-C₃N₄ composites are shown in Supplementary material, Figure S4. On the g-C₃N₄ curve, the wide absorption peak at 1,243–1,636 cm⁻¹ represents the specific stretching mode of C-N heterocyclic ring, and the peak at 808 cm⁻¹ may be attributed to the classical stretching mode of triazine units (Zhang et al. 2013). In addition, broadband in the range of 3,000–3,500 cm⁻¹ shows the elastic oscillation modes of water molecules on the N–H and O–H outer surfaces. The ZnS/g-C₃N₄ composite has similar spectra with pure g-C₃N₄, indicating that ZnS has no significant effect on the structure of g-C₃N₄.

The performance of CDI is significantly influenced by the electrode material's wettability, which can be determined using contact angle testing. Supplementary material, Figure S5 illustrates the contact angles of g-C₃N₄ and ZnS/g-C₃N₄ with water. The contact angle is typically less than 90° for hydrophilic materials and more than 90° for hydrophobic ones. Contact angles of 51.37° for g-C₃N₄ and 49.55° for ZnS/g-C₃N₄ show that both substances are hydrophilic. Furthermore, the composite was successfully constructed, increasing its wettability, as evidenced by the decrease in the ZnS/g-C₃N₄ contact angle. The performance of the desalination process is improved by encouraging contact between the NaCl solution and the electrode.

The EIS test was employed to evaluate the electrical resistance and results are shown in Figure 3(d). ZnS/g-C₃N₄ composite electrode showed a smaller charge transfer resistance. In addition, the slope in the low frequency part of the ZnS/g-C₃N₄ composite electrode is higher than that of g-C₃N₄, suggesting better ion diffusion.

Table 2 depicts the electrodes' desalination performance in various modes with an initial concentration of 200 mg/L NaCl solution, a flow rate of 5 mL/min, and a voltage of 1.2 V. The adsorption capacities of the g-C₃N₄//g-C₃N₄ (CDI), ZnS/g-C₃N₄//ZnS/g-C₃N₄ (CDI), g-C₃N₄//g-C3N4 (MCDI) and ZnS/g-C₃N₄///ZnS/g-C₃N₄ (MCDI) processes had adsorption capacities of 4.61, 8.38, 11.05, and 27.65 mg/g, respectively. The best desalination performance was achieved with the ZnS/g-C₃N₄ composite electrode in MCDI mode. It is clear that the use of membranes increases the electrode's adsorption capacity. The use of a membrane during the adsorption process can suppress the co-ion effect and allow the adsorbed ions to reach the adsorption maximum, which can adsorb more ions than CDI cell (Broseus et al. 2009). During desorption, the ion concentration on the surface of the electrode plate of the MCDI cell is higher than that of the CDI cell. At the same time, a large number of ions in the CDI cell's solution suffer from the disadvantage of attraction to the side electrode, causing the ion desorption of the MCDI cell's electrode to be greater than that of the CDI cell. As a result, given the effect of membrane addition on treatment capacity and desalination, the MCDI mode is appropriate.

To investigate the desalination performance of ZnS/g-C₃N₄ composites at different inlet flow rates. At an initial NaCl concentration of 200 mg/L and an applied voltage of 1.2 V, different inlet flow rates (5, 10, 15, and 20 mL/min) were chosen for this experiment. Figure 5(c) and (d) shows the variation of the conductivity of the NaCl solution with time, as well as the calculated electrosorption capacity of the electrode after reaching adsorption equilibrium. The adsorption capacity decreases gradually as the inlet flow rate increases. The electrosorption capacity was 27.65 mg/g at a flow rate of 5 mL/min, and the desalination effect was relatively good. This could be due to the short retention time at high flow rates, which could reduce ion adsorption (Tang et al. 2015). Increased solution inlet flow rate may result in a shorter retention time of the raw aqueous solution in the flow channel, and the ions migrating to the electrode surface may not have been adequately adsorbed and stored in the double electric layer (Agartan et al. 2019). Alternatively, the ions adsorbed on the electrode surface are less than the impact force applied and flow out of the desalination unit with the solution, reducing the electrode adsorption capacity.

In this study, a ZnS/g-C₃N₄ composite was successfully prepared by a one-step high-temperature calcination method and used as the main active material for electrodes in CDI desalination. SEM and TEM images revealed that spherical ZnS particles were evenly distributed on the layered structure of porous g-C₃N₄. N₂ adsorption–desorption tests and pore size distribution analysis demonstrated the ZnS/g-C₃N₄ composite had a larger specific surface area and suitable porous structure compared to g-C₃N₄ alone. Furthermore, ZnS/g-C₃N₄ exhibited favorable specific capacitance characteristics, relatively low resistance and high electrochemical stability. Notably, the ZnS/g-C₃N₄//ZnS/g-C₃N₄ (MCDI) electrodes showed a high adsorption capacity of 27.65 mg/g with an applied voltage of 1.2 V, a flow rate of 5 mL/min, and an initial NaCl concentration of 200 mg/L. When the initial concentration of NaCl increased to 400 and 600 mg/L, the adsorption capacity reached 50.26 and 65.34 mg/g, respectively. In summary, the ZnS/g-C₃N₄ composite electrodes could enable efficient desalination in MCDI systems.

The research is financially supported by the fundamental research program of Shanxi Province (No. 20210302123051).

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

The authors declare there is no conflict.