Enhanced Separation Performance of Graphene Oxide Membrane through Modification with Graphitic Carbon Nitride

By inergency On Mar 27, 2024

1. Introduction

Tritium (T) is a radioactive isotope of hydrogen that can enter the human body through inhalation and skin absorption, posing internal radiation hazards [1]. It occurs in minute quantities in nature and is primarily generated through nuclear reactions, particularly in heavy water reactors (HWRs) where heavy water is used for moderation and cooling purposes [2]. The current primary approach for tritium management in nuclear power plants involves its controlled release into the environment, such as discharging into the sea after dilution with a significant volume of seawater [3].

In the foreseeable future, the expansion of the nuclear industry presents significant challenges regarding the management and treatment of tritium-containing nuclear wastewater. Therefore, the application of tritium extraction or tritium separation technology is imperative in the treatment of nuclear wastewater containing tritium. However, the existing tritium separation techniques such as cryogenic distillation [4] and combined electrolysis catalytic exchange [5,6] are relatively energy-intensive, complex, and unfriendly to the environment. On the contrary, the membrane separation process exhibits significant potential for tritium treatment owing to its low energy consumption, exceptional selectivity, and high operational efficiency. Moreover, it has already gained widespread application in the field of industrial wastewater treatment [7]. As a result, it is necessary to explore an appropriate separation membrane and analyze its working principle and feasibility before designing a complete industrial facility.

Two-dimensional materials have gained significant attention from the public since the discovery of graphene [8]. Graphene’s unique atomic-level thickness, high mechanical strength, stable physical and chemical properties, and impermeability caused by dense atomic arrangement make its derivatives worthy of attention as membrane separation materials, such as porous graphene and graphene oxide [9,10]. In the past decade, numerous researchers have dedicated their efforts to investigating the preparation methods [11], mass transfer mechanisms [12], and separation performance [13,14] of graphene-based two-dimensional separation membranes. These endeavors have greatly advanced scientific progress in this field.

Porous graphene can be prepared by focused ion beam and can be used for hydrogen purification [15,16]. However, the filtration performance of this membrane entirely depends on the pore size, and since the kinetic diameters of different hydrogen isotope molecules vary little, and porous graphene films are not suitable for hydrogen isotope separation. Graphene oxide has become the most promising material for two-dimensional hydrogen isotope separation membranes due to its strong mechanical strength, excellent purification and separation performance, and suitability for large-scale production [17]. Laminar GO membrane contains hydroxyl and epoxy groups that maintain a greater spacing between the graphene layers in the unoxidized regions of the film [18]. Moreover, these groups tend to cluster together, creating a large network of unoxidized capillary channels for particles to pass through. Due to the unique hydrophilicity of oxygen-containing groups and capillary pressure generated by graphene’s hydrophobic layer, water molecules can rapidly penetrate through the capillaries network, achieving a perfect GO membrane penetration of water [19].

Currently, extensive research has been conducted to investigate the transport mechanism of gas, water, and solution ions across the GO membrane [20,21], as well as the water molecular arrangement between the interfaces of two graphene layers, particularly within the capillaries in the GO membrane [22]. Water is transported in the form of double-layer ice in the unoxidized region between the two nanosheets in the GO film, that is, inside the nanoscale capillary [23]. It is destroyed (melted) as it passes through the oxygen-containing groups at the edge of the nanosheet, after which the water molecules penetrate into the next capillary layer and recombine into a double-layer ice structure to continue its transport [24]. The GO membrane exhibits a remarkable ion screening capability in solution desalination experiments that smaller ions (with a hydration radius of less than 4.5 A) can rapidly permeate through the membrane, while larger and organic molecules are virtually impermeable [18].

The hydrophilic nature of the GO membrane makes it a promising candidate for isotope water separation technology. An experiment proved that the content of heavy isotopes (deuterium and tritium) would be significantly reduced after hydrogen isotope water passed through the GO membrane [25]. In recent years, a large number of researchers have invested in the study of the optimization of the separation performance of GO membranes. Compared to GO membranes, various separation membranes optimized based on GO materials show better separation properties, such as higher penetration rates or higher separation factors. These modification methods of the GO membrane include the reduction method [26], surface modification method [27], intercalation method [28], and so on. RGO membranes exhibited a comparable D₂O blocking rate to GO membranes while demonstrating superior mechanical stability and serving as an effective material for the separation of hydrogen isotope water mixtures [29]. A research team employed perfluoroalkyl substances [30] or fluorinated silica nanoparticles [31] to modify one side of the GO membrane forming a hydrophobic or super-hydrophobic surface, resulting in modified GO membranes exhibiting enhanced separation factors.

It is worth mentioning that researchers commonly employ heavy water instead of tritium to analyze the separation of hydrogen isotopes in water, primarily due to the potential radiation hazards involved. The H/D separation factors can reflect the H/T separation factors with some theoretical calculations [25].

In this study, GO membranes were modified with graphitic-phase carbon nitride (g-C₃N₄) and tested using self-made distillation equipment to demonstrate that this intercalation method can enhance the performance of the GO membranes in separating hydrogen isotopes. The impact of intercalating different quantities of g-C₃N₄ on the selective separation of GO membranes was also investigated.

2. Materials and Methods

2.1. Materials

The nature graphite powder was purchased from Tianjin Damao Co., Ltd. (Tianjin, China). The g-C₃N₄ dispersion was supplied by Nanjing XFNANO Technology Co., Ltd. (Nanjing, China). Potassium permanganate (KMnO₄), concentrated sulfuric acid (H₂SO₄, 98%), hydrogen peroxide (H₂O₂, 30%), and hydrochloric acid (HCl, 37%) were supplied from Shanghai Aladdin Biochemical Technology Co., Ltd. (Shanghai, China). PC (polycarbonate) membranes and AAO (anodic aluminum oxide) membranes with 0.22 μm pore size were obtained from Whatman Co., Maidstone, UK. Hydrophobic PTFE membranes (tetrafluoroethylene, ϕ 47 mm, pore size ~1 μm, Millipore Co., Billerica, MA, USA) were used as the support in the distillation tests. The heavy water (99.9 at. % D) was purchased from Energy Chemical Co., Ltd. (Anqing, China). Deionized water (18.25 MΩ) was produced by Ultra-pure Water Purifiers (Ulupure UPR-Ⅱ, Chengdu, China) in the lab.

2.2. Fabrication of Membranes

The GO nanosheets were synthesized from graphite powder using the improved Hummers method [32]. GO dispersion solution was prepared by the ultrasonic dispersing of dry GO in deionized water. The GO membrane was fabricated on a PC substrate through vacuum filtration. In this study, 0.2 g of dried GO solid was sonicated in 100 mL of deionized water to make 2 mg/mL of GO dispersion. After that, 5 mL, 7.5 mL, 10 mL, and 15 mL GO dispersion solutions were taken and prepared into GO membranes of varying thicknesses by vacuum filtration.

The preparation of the GO/g-C₃N₄ composite membrane followed a similar procedure as that of the GO membrane. The mass concentrations of GO and g-C₃N₄ suspensions were 2 mg/mL and 0.18 mg/mL, respectively. In this work, 2 mL, 4 mL, 6 mL, 8 mL, and 10 mL g-C₃N₄ suspensions were added into 5 mL GO suspensions. The suspension of GO and g-C₃N₄ was intensively mixed by ultrasonication for 2 h and then filtered on the AAO substrate to form a membrane. Following a 24 h drying period at room temperature, the independent membrane could be easily detached from the substrate.

2.3. Characterization of the Membranes

The surface and cross-section morphology of membranes were studied by a field emission scanning electron microscope (FE-SEM, JEOL JSM-7900F, Yamagata, Japan). The gold-coated membranes were soaked in liquid nitrogen to make them easier to tear. The observation of GO nanosheets was supplied by a tunneling electron microscope (TEM, Hitachi H-7650, Marunouchi, Japan). X-ray diffraction (XRD, Bruker D8 ADVANCE, Karlsruhe, Germany) equipped with Cu Kα radiation (λ = 0.154 nm) was applied to measure the d-spacing of the prepared membranes. X-ray photoelectron spectroscopy (XPS, Thermo Scientific ESCALAB QXi, Morecambe, UK) and Raman spectra (Horiba Jobin Yvon LabRam HR800, Palaiseau, France) were used to recognize the chemical composition of the membrane surface.

2.4. Selective Separation Tests

2.4.1. Static Diffusion Experiment

The static diffusion experiment used a 50 mL glass sample bottle, which was filled with 20 mL H₂O or D₂O (99.9%) liquid. The GO membrane was covered at the opening of the sample bottle, and the place where the mouth of the bottle contacted the membrane was sealed using an O-type silicone ring. The experiment was carried out at room temperature, and the weight of bottles was weighed every 24 h, based on which the amount of liquid lost was calculated. The experiment was conducted for a total of 43 days. The permeate flux (J) was calculated by equation:
where m, A, and t represent the mass loss of liquid in the bottle, the membrane area, and the experiment time.

2.4.2. Distillation Tests

To analyze the differences in the performance of GO membranes with different properties in the separation of hydrogen isotope water, a custom-made experimental device for the separation of hydrogen isotope water was established. As shown in Figure S1, the device is divided into three sections: the feed section consists of an evaporation chamber with a capacity of approximately 100 mL and a pressure-reducing valve; the middle section comprises a membrane carrier; and the collect section is connected to a sample bottle for collecting the filtered liquid.

In the test, the evaporation chamber was loaded into a resistance oven and insulated with thermal insulation cotton. All stainless steel components, except for the evaporation chamber and pressure-reducing valve, were covered with a glass fiber heating sleeve to prevent the condensation of water vapor and heavy water vapor within the apparatus. Two temperature controllers were used to regulate the temperature of the resistance furnace and heating sleeve, respectively. A sample collection bottle was positioned on the collection side and connected to the stainless steel tube via a high-temperature-resistant silicone hose. The GO membrane was sealed in the middle section between the silicone and stainless steel gaskets to make sure the whole device was airtight during the process. Considering the vapor pressure on the feed side can reach 0.14 MPa at 110 °C, a support PTFE membrane was attached to the permeate side of the GO membrane to prevent it from being ruptured by the high pressure.

The molar concentration of heavy water in the chamber of 10% was achieved by mixing pure D₂O with deionized water. The deuterium and hydrogen abundance of the permeated water were measured by NMR (Bruker AVANCE III 400 MHz, Karlsruhe, Germany). The heavy water samples, each with concentrations of 6%, 8%, 10%, 12%, and 14%, were combined in a 1:1 ratio with pure acetone solution. A small quantity of the resulting mixture was then transferred into an NMR tube for analysis. Two distinct peaks would appear in the ¹H NMR spectrum, namely the water peak caused by the vibrations of hydrogen in water and the acetone peak caused by vibrations of hydrogen in methyl groups in acetone. Figure S2 shows a ¹H NMR spectroscopy of a mixture of 10% heavy water and acetone with a 1:1 volume ratio. The left peak represents the water peak and the right peak represents the acetone peak, respectively. The area ratio of the two peaks is I_H2O/I_acetone = 1.182. The area ratio of the two peaks was positively correlated with their contents. Finally, a graphical representation illustrating the relationship between the concentration of heavy water and the corresponding ratio is presented, revealing a linear correlation upon fitting, thus establishing a standard curve, as shown in Figure S3. To obtain the concentrations of heavy water in distillation tests, the unknown concentration of heavy water to be analyzed was mixed with an acetone solution in a 1:1 ratio, and a portion of the resulting mixture was introduced into the NMR tube. By comparing the peak area ratio of the two observed peaks in the test results with a standard curve, the heavy water concentration of the sample under investigation was determined.

The separation factor q can be calculated by equation:

$q = \frac{}{} C_{H}^{'} / C_{D}^{'}$

(2)

where C_H and C_D represent the hydrogen and deuterium abundance of permeated water, and C_H′ and C_D′ denote the hydrogen and deuterium abundance of feed water. The deuterium abundance C_D is equal to the concentration of heavy water which is given by NMR results and the hydrogen abundance is equal to 1-C_D.

4. Discussion

The separation factors of the GO/g-C₃N₄ membranes were similar to the GO membrane. The collective migration of water molecules in the GO membrane and the resulting hydrogen isotope separation performance are attributed to the unique structure of the GO membrane. The oxygen-containing groups located at the edges of GO nanosheets provide support for interlayer spacing, allowing water molecules to pass through and enter subsequent layers via structural defects within or at the edge of the membrane [36]. Due to its triazine or heptazine structure [37], two-dimensional g-C₃N₄ possesses an ample number of molecular-level pores for water molecules to pass through, while amino groups are present at the edges, resembling oxygen-containing groups in GO and capable of forming hydrogen bonds with water. The chemical properties of carbon nitride nanosheets closely resemble those of GO nanosheets, thereby preserving the separation efficiency of GO membranes towards hydrogen isotope water.

The permeate flux of the composite membranes exhibited an obvious increase after being modified with g-C₃N₄. Three contributing factors are postulated to account for this result:

Large barely stripped g-C₃N₄ bulks are first deposited at the bottom of the membrane in the vacuum filtration process, as shown in Figure 1d. These particles contribute to an increased surface roughness of the membrane. A study has demonstrated that the high permeability of GO membranes observed may be attributed to the disordered microstructure of the membrane [38];
Smaller particles of partially stripped g-C₃N₄ with a transverse size of 10 μm and height of 3 μm can be observed to be inserted between the laminates, as illustrated in Figure 1f. The I_D/I_G ratio of the GO/g-C₃N₄ membrane is larger than the GO membrane in the Raman analysis, indicating that the interlayer particles disrupt the originally relatively complete structure and result in more defects within the composite membrane.
The shift of the peak in the XRD pattern can prove that single or fewer layers of g-C₃N₄ nanosheets are inserted between the layers of GO nanosheets and the layer d-spacing of the GO membrane is modified [37]. The increase in d-spacing from 7.6 Å to 7.9 Å contributes to the higher permeability [39].

Moreover, the static diffusion test was carried out at room temperature for 43 days, during which the permeation rate of the membranes was maintained as time went by. In the distillation experiment, the temperature of the film is maintained at 110 °C under the action of the glass fiber heating jacket. At this temperature, the permeability rate of the GO membrane will decrease with the increase in time [29], which may be attributed to the membrane being reduced at high temperatures. This results in a slightly lower permeation rate of the GO membrane in the distillation experiment than in the static diffusion experiment performed at room temperature.

The outstanding performance of the GO/g-C₃N₄ membrane in heavy water separation suggests its potential application in the treatment of tritiated wastewater in the future nuclear industry.