An amino-functionalized graphene oxide membrane, a preparation method, application and separation performance prediction method thereof
Patent Information
- Application Number
- CN202610617354.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-05-07
- Publication Date
- 2026-10-23
AI Technical Summary
尽管这两种电荷调控策略均已被证明能够实现有效的一/二价阳离子分离，但目前尚未形成共识，即哪一种方法更具优势与效率
[0059]本发明的有益效果是：将分子动力学（MD）模拟与实验相结合，提出正电荷改性在二维GO膜中对一/二价离子分离更具优势。本研究采用MD模拟，在原子尺度上系统对比了正电（氨基改性）和负电（羧基功能化）GO膜中K+/Mg2+跨膜传输的微观机制。在负电GO膜中，静电吸引诱导离子在膜表面富集，触发离子间库仑排斥并产生knock-on效应，依次推动离子跨膜通道传输。相反，在正电GO膜中，Mg2+因其较高价态产生的强静电排斥而被有效截留，形成高传输势垒，而K+实现选择性传输。基于模拟结果，进一步通过简便的原位改性策略制备了正电（氨基化）GO膜。该正电膜在pH 3-11范围内保持良好的离子选择性，在pH = 3时达到最佳性能，选择性为77，渗透通量为0.26 mol m-² h-¹。
Smart Images
Figure CN122955366A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to an amino-functionalized graphene oxide membrane, its preparation method, application, and separation performance prediction method, belonging to the field of membrane separation technology. Background Technology
[0002] Selective transport of monovalent / divalent cations is crucial in water treatment, resource recovery, and energy storage. To achieve efficient and precise sieving of monovalent / divalent cations, researchers have developed various membrane control strategies, including size sieving, specific recognition, and surface charge modulation. Among these, charge modulation strategies, which utilize electrostatic differences to regulate ion transport, are considered an effective method to overcome the "trade-off" problem in membrane separation.
[0003] Two-dimensional materials (such as graphene oxide, GO) are considered highly promising membrane materials due to their single-atom-layer thickness. The GO surface is rich in various oxygen-containing functional groups (such as hydroxyl, epoxy, and carboxyl groups), which not only endow it with intrinsically tunable localized negative charge properties but also provide reaction sites for chemical modification. To date, both negatively and positively charged GO membranes have been used for the separation of monovalent / divalent cations. Although both charge modulation strategies have been proven to achieve efficient monovalent / divalent cation separation, there is currently no consensus on which method is more advantageous and efficient. Especially in nanoscale confinement, the mechanisms by which electrostatic attraction and repulsion regulate the transmembrane transport pathways and kinetics of monovalent / divalent cations at the microscopic level remain not fully understood. This gap in understanding limits the rational design and controllable construction of high-performance charge-controlled GO membranes. Summary of the Invention
[0004] This invention modifies graphene oxide films by amino-functionalization and combines this with thermal reduction to regulate the film structure and surface charge properties, enabling the resulting GO film to effectively promote K+. + Selective permeation and inhibition of Mg 2+ Through, thus achieving K + / Mg 2+ Highly efficient separation. The membrane preparation method is simple, employing in-situ modification with ammonia and vacuum filtration, requiring no exogenous polymer modifiers and easy to implement. The resulting membrane maintains its separation capability within a pH range of 3–11, exhibiting superior selectivity and flux under acidic conditions, and demonstrating good operational stability. Furthermore, this invention incorporates molecular dynamics simulation prediction methods, which can guide membrane material design from the perspectives of ion permeation, interaction energy, and transport barriers, providing a basis for the rational construction of charged two-dimensional membranes and the optimization of ion separation performance.
[0005] A type of K + / Mg 2+The separated GO membrane includes a porous support layer and a graphene oxide separation layer disposed on the surface of the porous support layer, wherein the graphene oxide separation layer is an amino-functionalized graphene oxide layer.
[0006] The porous support layer is a nylon filter membrane.
[0007] The amino-functionalized graphene oxide layer is formed by loading ammonia-modified graphene oxide onto a porous support layer.
[0008] The graphene oxide separation layer contains amino functional groups corresponding to CN bonds.
[0009] The GO membrane exhibits a positive charge on its surface under acidic conditions.
[0010] The method for preparing a GO membrane includes the following steps:
[0011] (1) Ammonia water was added to the graphene oxide dispersion to react and obtain amino-functionalized graphene oxide;
[0012] (2) The amino-functionalized graphene oxide is dispersed in water to obtain a dispersion;
[0013] (3) The dispersion is loaded onto the surface of the porous support layer to form a graphene oxide separation layer;
[0014] (4) The obtained membrane is thermally reduced to obtain the GO membrane.
[0015] In step (1), the concentration of the graphene oxide dispersion is 0.1-50 mg / g.
[0016] In step (1), the ratio of ammonia water to graphene oxide dispersion is 0.5:(20-100).
[0017] The reaction temperature in step (1) is 40-120℃ and the reaction time is 2-20h.
[0018] Step (1) is followed by a step of centrifugation to collect the solid product.
[0019] In step (2), the amount of amino-functionalized graphene oxide added to deionized water is controlled at a mass-to-volume ratio of 0.5 mg:(100-500) mL, and ultrasonic dispersion is performed; and the loading of amino-functionalized graphene oxide on the porous support layer surface is 0.01-0.1 mg / cm³. 2 .
[0020] The heat treatment temperature in step (4) is 120 to 200°C.
[0021] The GO film in K + / Mg2+ Applications in separation.
[0022] In the application K + Mg preferentially permeates through the GO membrane. 2+ It was selectively withheld.
[0023] The application was carried out in the pH range of 3 to 11.
[0024] A method for predicting the K-axis of charged graphene oxide films + / Mg 2+ Molecular dynamics simulation methods for separation performance include the following steps:
[0025] (1) Establish a nanochannel model of graphene oxide membrane, wherein the graphene oxide membrane is an amino-functionalized graphene oxide membrane or a carboxyl-functionalized graphene oxide membrane.
[0026] (2) Water-containing chambers are set on both sides of the nanochannel model, and K is set in the feed-side chamber. + Mg 2+ And Cl⁻ used to maintain electroneutrality;
[0027] (3) Apply external pressure to the nanochannel model and perform non-equilibrium molecular dynamics simulations to obtain K. + and Mg 2+ Transmembrane transport trajectory;
[0028] (4) Calculate K based on the simulation results + and Mg 2+ The number of ions permeated was determined, and the interaction energy between ions and the graphene oxide membrane and / or the average force potential for ion transport across the membrane were calculated.
[0029] (5) According to K + With Mg 2+ The differences in the amount of permeate, the differences in the interaction energy between ions and the graphene oxide membrane, and / or the differences in the ion transmembrane transport barrier are used to predict the effect of the charged graphene oxide membrane on K. + / Mg 2+ Separation performance.
[0030] In step (1), the graphene oxide membrane nanochannel model is constructed based on the Lerf-Klinowski model.
[0031] In step (1), the graphene oxide model has dimensions of 25-35 Å in both the x and y directions.
[0032] In step (1), amino functional groups or carboxyl functional groups are uniformly distributed at the edges of the graphene oxide sheets.
[0033] In step (1), the lateral distance of the graphene oxide sheet is determined by umbrella-shaped sampling simulation, and a single-layer graphene oxide membrane channel model is constructed with a distance of 0.8 nm based on the average force potential curve.
[0034] In step (2), a layer of rigid pristine graphene is placed on each side of the nanochannel model as a piston to apply pressure and maintain structural stability.
[0035] A 90 Å vacuum layer was set in the z-direction to eliminate the interaction between the system and its periodic mirror image.
[0036] The molecular dynamics simulation applied periodic boundary conditions in the three Cartesian directions (x, y, and z), and the water molecules were described using the SPC / E model.
[0037] The molecular dynamics simulations used the Amber force field to describe the interatomic interactions. The nonbonded interactions between atoms were calculated using the Lennard-Jones 12-6 potential and the Coulomb potential. The cross-interaction parameters were derived using the Lorentz-Berthelot mixing rule.
[0038] All sp² hybridized carbon atoms in graphene oxide and graphene sheets are considered electrically neutral, and the partial charges of carboxyl and amino groups are determined using the RESP2 method.
[0039] To maintain the rigidity of the graphene piston, a dihedral torsional barrier is applied to it.
[0040] In the nonequilibrium molecular dynamics simulation, the cutoff distance for van der Waals interactions is 0.5-2 nm, and long-range electrostatic interactions are handled using the particle mesh Ewald method.
[0041] The simulation process includes first minimizing energy, followed by a 60-70 ns simulation under a canonical ensemble.
[0042] During the simulation, the carbon sheets of the graphene oxide film were positionally restricted, while the oxygen-containing functional groups remained relaxed.
[0043] The system temperature was maintained at 295-300K using a V-rescale thermostat. The initial velocity was randomly generated, the integration time step was 0.5-1 fs, and the trajectory coordinates were recorded every 2-8 ps.
[0044] The external pressure applied in step (3) is 180-220 MPa, and this external pressure is achieved by applying a force to each carbon atom on the original graphene piston, the force being calculated by the following formula:
[0045]
[0046] Where f is the force applied to each carbon atom, P is the pressure, A is the piston area, and n is the total number of carbon atoms in the piston.
[0047] In step (4), the interaction energy between the ions and the graphene oxide film is calculated using the following formula:
[0048]
[0049] in, Represents the energy of ion-GOGO compounds. and These correspond to the energies of isolated ions and isolated GO nanosheets, respectively.
[0050] The more negative the value, the stronger the attraction between the ions and the graphene oxide film; The more positive the value, the stronger the repulsion between the ions and the graphene oxide film.
[0051] In step (4), the energy barrier for ion transmembrane transport is calculated using the average force potential, which is calculated using the adaptive bias force method implemented by the Colvars module in GROMACS.
[0052] In step (4), the radial distribution function between the ion and the functional group is also calculated to characterize the distribution probability of the ion near the functional group.
[0053] The cutoff distance of the radial distribution function is set to half the length of the simulated box.
[0054] In step (4), the average residence time of ions in the functional group range of 4-6 Å is also counted to characterize the adsorption and enrichment behavior of ions near the functional group.
[0055] In step (4), the transmembrane transport trajectory of ions is also analyzed to identify the inter-ion coulombic repulsion and sequential knock-on transport behavior triggered by the enrichment of ions near the carboxyl group.
[0056] In step (5), when the amino-functionalized graphene oxide film reacts with Mg... 2+ Showing higher than K + When the transmembrane transport barrier is given, it is predicted that the membrane has K0. + Selective transmission performance.
[0057] In step (5), when the carboxyl-functionalized graphene oxide film is subjected to K + and Mg 2+ When exhibiting similar permeation rates, the membrane is predicted to be effective against K. + / Mg 2+ The selectivity is low.
[0058] Based on the prediction results of step (5), amino-functionalized graphene oxide film was selected as K. + / Mg 2+ Candidate membranes for separation.
[0059] The beneficial effect of this invention is that it combines molecular dynamics (MD) simulations with experiments, proposing that positively charged modification is more advantageous for the separation of monovalent / divalent ions in two-dimensional GO membranes. This study uses MD simulations to systematically compare the K+ concentrations in positively charged (amino-modified) and negatively charged (carboxyl-functionalized) GO membranes at the atomic scale. + / Mg 2+ The microscopic mechanism of transmembrane transport. In negatively charged GO membranes, electrostatic attraction induces ion enrichment on the membrane surface, triggering Coulomb repulsion between ions and generating a knock-on effect, which in turn drives ion transport across the membrane channel. Conversely, in positively charged GO membranes, Mg... 2+ It is effectively trapped due to the strong electrostatic repulsion generated by its higher valence state, forming a high transport barrier, while K + Selective transport was achieved. Based on simulation results, a positively charged (aminated) GO membrane was further prepared using a simple in-situ modification strategy. This positively charged membrane maintained good ion selectivity in the pH range of 3-11, achieving optimal performance at pH = 3 with a selectivity of 77 and a permeation flux of 0.26 mol m - ² h - ¹. Attached Figure Description
[0060] Figure 1 Schematic diagram of monovalent / divalent cation separation in graphene oxide film with positive and negative charge regions.
[0061] Figure 2 (a) Protonated PGO nanosheet model; (b) Protonated NGO nanosheet model.
[0062] Figure 3 : Average potential energy distribution curve of two PGO / NGO nanosheets approaching each other in aqueous solution.
[0063] Figure 4 Aminofunctionalization and thermal reduction process of GO membranes.
[0064] Figure 5 (a) Average residence time of ions near functional groups; (b) Energy barrier for ion transport across membranes.
[0065] Figure 6 (a) A snapshot of the ion distribution in the PGO membrane; (b) A snapshot of the ion distribution in the NGO membrane. Yellow and purple spheres represent K⁺ and Mg²⁺, respectively. Black arrows indicate the knock-on effect in the PGO membrane.
[0066] Figure 7 (a) Average residence time of ions near functional groups; (b) Radial distribution function (RDF) of ions and functional groups.
[0067] Figure 8 : Average potential energy distribution curve of ion transmembrane transport.
[0068] Figure 9 (a) FTIR spectrum of the NGO-X membrane (X represents the reaction temperature, unit: °C); (b) T r FTIR spectrum of NGO-S membrane (S represents thermal reduction temperature, unit: °C).
[0069] Figure 10 (a) XPS C1s spectrum of PGO membrane; (b) XPS C1s spectrum of NGO-60 membrane; (c) T r XPS C 1s spectrum of NGO-160 membrane; (d) XPS N 1s spectrum of PGO membrane; (e) XPS N 1s spectrum of NGO-60 membrane; (f) T r XPS N 1s spectrum of NGO-160 membrane.
[0070] Figure 11 (a) Zeta potential of PGO and NGO-X membranes (X represents reaction temperature, unit: °C); (b) T r zeta potential of NGO-S membrane (S represents thermal reduction temperature, unit: °C)
[0071] Figure 12 (a, c) SEM images of the NGO-60 membrane; (b, d) T r SEM image of the NGO-160 membrane.
[0072] Figure 13 : (a)T r Raman spectra of NGO-S films; (b) T r XRD diffraction pattern of NGO-S film; (c)T r Water contact angle of NGO-S membrane; (d)T r K⁺ permeability and K⁺ / Mg²⁺ selectivity of NGO-S membrane in a 0.2 M single-salt system. S represents thermal reduction temperature in °C.
[0073] Figure 14 :(a)T r (a) Zeta potential of NGO-160 membrane; (b) K⁺ permeability and K⁺ / Mg²⁺ selectivity in 0.2 M mixed salt solution under different pH conditions; (c) Performance comparison of ion permeability and selectivity; (d) Positive charge T at pH = 3.r Stability testing of NGO-160 membrane. Detailed Implementation
[0074] The following raw materials were used in the preparation of the GO membrane: monolayer graphene oxide (GO) dispersion (lateral size: 5–8 μm, concentration: 10 mg / g, monolayer rate >99.5%), and ammonia water (25-28 wt%).
[0075] 0.5 mL of ammonia solution was added to 60 mL of graphene oxide dispersion (the raw material was diluted to a concentration of 0.5 mg / g), and the reaction was carried out at 40, 60, 80, 100, and 120 °C for 6 h, respectively. After centrifugation to collect the solid product, 0.5 mg of powder was redispersed in 200 mL of deionized water and sonicated for 60 min to obtain a uniformly dispersed amino-functionalized graphene oxide suspension. Subsequently, the 200 mL suspension was vacuum filtered onto a nylon support membrane (diameter: 47 mm, pore size: 200 nm) at 2 bar pressure to prepare an amino-functionalized graphene oxide membrane. The resulting membrane was denoted as NGO-X, where X represents the reaction temperature. In addition, the above membrane was heat-treated at different temperatures (120, 140, 160, 180, and 200 °C), and the resulting membrane was denoted as T. r NGO-S, where S represents the heat treatment temperature.
[0076] The test methods for membrane separation performance are as follows:
[0077] Ion separation experiments were conducted using a U-shaped osmosis apparatus. The membrane under test was installed as a diaphragm between the two chambers, with the selective layer facing the feed side. In the single-salt experiments, KCl or MgCl2 solutions of different concentrations were introduced into the feed chamber. In the mixed-salt experiments, a mixed solution containing different concentrations of KCl and MgCl2 was used. Deionized water was introduced into the osmosis chamber. Continuous stirring was performed during the experiments to reduce concentration polarization. The ion permeation flux was determined using the following formula:
[0078]
[0079] In the formula, (mol m⁻² h⁻¹) represents the ion permeation flux. (L) represents the volume of the osmotic solution. (m²) represents the effective membrane area. (g mol⁻¹) is the molar mass. t (mg L⁻¹ h⁻¹) represents the rate of concentration change. Ion selectivity is calculated using the following formula:
[0080]
[0081] In the formula, S represents ion selectivity, and JS1 (mol m⁻² h⁻¹) is the osmotic flux of K⁺ ions. (mol m⁻² h⁻¹) represents the osmotic flux of Mg²⁺ ions. Single-salt flux and selectivity were measured using a conductivity meter, while mixed-salt flux and selectivity were determined by inductively coupled plasma atomic emission spectrometry.
[0082] Models and Methods
[0083] A GO structure model with dimensions of 30×30 Ų (x and y directions) was constructed based on the Lerf-Klinowski model. The oxidation level was set to 20%, and the ratio of hydroxyl to epoxy groups was 1:1. Amino or carboxyl functional groups were uniformly distributed along one edge of the GO sheet. Figure 2 To determine the slit distance, we performed umbrella sampling simulations (SI). Based on the average potential curve, a monolayer GO membrane channel model was constructed using a slit distance of 0.8 nm. Figure 2 A rigid layer of pristine graphene was placed on each side of the model as pistons to apply pressure and maintain structural stability. A 90 Å vacuum layer was set along the z-axis to eliminate the interaction between the system and its periodic mirror image. Both chambers were filled with water molecules. In the ion transport simulation, K0 in the feed chamber... + and Mg 2+ The initial concentration was set at 0.5 mol / L to ensure a sufficient ion supply. Chloride ions were added to maintain electroneutrality.
[0084] Molecular dynamics simulations were performed using GROMACS 2024.5. Periodic boundary conditions were applied in all three Cartesian directions (x, y, and z). Water molecules were described using the SPC / E (Extended Simple Point Charge) model. Trajectory visualization and configuration analysis were performed using VMD (Visual Molecular Dynamics). Intermolecular interactions were described using the Amber force field, where nonbonded interactions were calculated using the Lennard-Jones (LJ) 12-6 potential and the Coulomb potential, as expressed below:
[0085]
[0086] In the formula, Represents the depth of the potential energy trap. Indicates the equilibrium separation distance. The distance between atoms i and j. and These correspond to the partial charges of atoms i and j, respectively.
[0087] Interaction parameters were derived using the Lorentz-Berthelot mixing rule. The Amber force field has been widely used for the verification of carbon-based materials. In this simulation, this force field reliably preserved the layered structure of GO. All sp² hybridized carbon atoms in both GO and graphene sheets were considered electrically neutral. The partial charges of the GO functional groups were taken from the literature. The specific charges of the carboxyl and amino groups were determined using the RESP2 method, and the results are shown in Table 1.
[0088] Table 1 RESP2 Fitted Charges
[0089] C -0.237 C 0.364 C 0.915 N -0.481 O -0.837 H 0.372 O -0.840 H 0.371 - - H 0.374
[0090] To maintain the rigidity of the graphene piston, a high dihedral torsional barrier was applied—a method widely used in molecular simulations. For nonequilibrium molecular dynamics simulations, the cutoff distance for van der Waals (vdW) interactions was 1.0 nm, while long-range electrostatic interactions were handled using the particle mesh Ewald (PME) method. The interactions between ions and the GO film were calculated using the following formula:
[0091]
[0092] In the formula, Represents the energy of ion-GOGO compounds. and The energies correspond to isolated ions and isolated GO nanosheets, respectively. The more negative the value, the stronger the attraction; conversely, the more positive the value, the stronger the repulsion.
[0093] The simulation process began with energy minimization in the microcanonical ensemble (NVE), followed by a 65 ns simulation in the canonical ensemble (NVT). During this timescale, the feed chambers of all systems were water-free. To ensure statistical reliability, each system underwent three independent, repeated simulations. During the simulation, the carbon sheets of the GO membrane were positionally confined, while the oxygen-containing functional groups remained fully relaxed. The system temperature was maintained at 298.15 K using a V-rescale thermostat, and the initial velocity was randomly generated. Trajectory coordinates were recorded every 5 ps, with an integration time step of 1 fs. A high pressure of 200 MPa was applied to suppress thermodynamic fluctuations on the nanosecond scale and enhance the signal-to-noise ratio during transmembrane transport. This pressure was achieved by applying a force to each carbon atom on the pristine graphene piston. The pressure calculation expression is as follows:
[0094]
[0095] In the formula, f is the force, p is the pressure, A is the piston area, and n is the total number of carbon atoms in the piston. The average force potential (PMF) is calculated using the adaptive bias force (ABF) method implemented in the Colvars module of GROMACS.
[0096] The radial distribution function (RDF) provides a statistical description of the radial arrangement of particles in a system. The formula for calculating the RDF is as follows:
[0097]
[0098] In the formula, Represents the radial distribution function. Let be the density of particle i at a distance r around particle j. Let be the average density of particle i within a sphere of radius r centered at particle j. The cutoff distance is set to half the length of the box.
[0099] MD Simulation Ion Separation Performance and Mechanism
[0100] Ion transport behavior was studied using MD simulations. Figure 1 In the negatively charged region of the original graphene oxide membrane (PGO), K⁺ and Mg²⁺ exhibit similar permeability, with a selectivity of only 1.1. In contrast, in the positively charged region of the -NH⁺-grafted graphene oxide membrane (NGO), efficient ion sieving is achieved: Mg²⁺ is efficiently repelled, while the permeability of K⁺ is only slightly lower than in the negatively charged region. Figure 5 (a) To elucidate the microscopic mechanism underlying the performance difference between the two surface charge states of GO films, the interaction energy between ions and the film was calculated. The results show that the interaction energy between PGO and ions is negative, characterizing electrostatic attraction; while the interaction energy between NGO and ions is positive, characterizing electrostatic repulsion. Figure 5 (b) This aligns with the fundamental principle of "opposites attract, likes repel". Furthermore, due to the higher intrinsic charge density of Mg²⁺, its electrostatic interaction strength is significantly higher than that of K⁺.
[0101] The electrostatic attraction between K⁺ / Mg²⁺ and the -COO⁻ group leads to the adsorption and enrichment of ions near -COO⁻. Figure 6 (a). Conversely, due to electrostatic repulsion, no ion aggregation was observed near the -NH3⁺ group in the NGO membrane (a). Figure 6 (b). This phenomenon was quantitatively characterized by the residence time of ions in the functional group within a 5 Å range and the ion-functional group radial distribution function (RDF). Due to favorable adsorption, the residence time of ions near -COO⁻ is much longer than that near -NH3⁺. Figure 7(a). Meanwhile, the RDF peak of the -COO⁻ ion interaction is significantly stronger than that of the -NH⁺ ion interaction, further confirming that due to electrostatic attraction, ions have a higher distribution probability around -COO⁻. Figure 7 (b). Further analysis of the transport trajectory shows that ion enrichment driven by electrostatic attraction induces Coulomb repulsion between ions, thereby sequentially "breaking down" and driving ion transport across the membrane. Figure 6 (a). This "knock-on effect" is an important mechanism for achieving rapid and selective ion transport in biological ion channels and artificial supramolecular channels. However, the negatively charged regions of GO membranes lack specific ion recognition sites similar to those in biological systems, ultimately leading to near-nonselective ion transport.
[0102] To further analyze the problem from an energy perspective, the average potential energy (PMF) of ion transmembrane transport was calculated. Figure 8 Electrostatic attraction between ions and the -COO⁻ group induces ion enrichment, followed by sequential knock-on transport driven by the repulsion between the enriched ions. This significantly reduces the transmembrane energy barrier difference between K⁺ and Mg²⁺ when passing through the PGO membrane. Furthermore, oxygen in the -COO⁻ group can partially replace oxygen in water molecules, thus compensating for the energy required for ion dehydration and further reducing the transport energy barrier. In contrast, in the NGO membrane, ions need to overcome electrostatic repulsion from the -NH3⁺ group, facing a higher transport energy barrier. The larger energy barrier difference between K⁺ and Mg²⁺ is beneficial for selective transport.
[0103] Membrane preparation and characterization
[0104] GO membranes with varying deposition amounts were prepared by vacuum filtration to optimize membrane thickness and obtain a defect-free separation layer. Cross-sectional SEM characterization showed that membrane thickness increased with increasing deposition amount. Single-salt permeation experiments indicated that at the optimized deposition amount of 0.075 mg / cm², the membrane achieved the highest ion selectivity while maintaining moderate permeability. To balance the trade-off between permeability and selectivity, this deposition amount was determined as the optimal preparation parameter for subsequent modification studies.
[0105] The amino functional group was introduced into GO by a simple reaction in ammonia water at different temperatures. Figure 4 FTIR absorption spectroscopy showed that, compared with the original GO film, the NGO film exhibited a characteristic C–N stretching vibration peak at approximately 1260 cm⁻¹ at all four reaction temperatures. Figure 9 (a) confirmed the successful introduction of the amino functional group. XPS characterization further verified the successful grafting of the amino group through the appearance of the C–N peaks. Figure 10The results also showed that the nitrogen content in the NGO membrane gradually increased with increasing reaction temperature, from 2.46% of the original GO to 5.55% at 100 °C. Figure 10 ).
[0106] The effect of amination temperature on the charge state of the GO film was characterized by zeta potential testing. Figure 11 As shown in Figure a, the original GO film exhibits a negative charge. When the reaction temperature exceeds 60 °C, the GO film undergoes a charge transition and becomes positively charged. This phenomenon may originate from the partial thermal decomposition of negatively charged oxygen-containing functional groups, which can be observed by the decrease in oxygen content in the XPS spectrum. Figure 10 The increase in the water contact angle of the GO membrane (ab) is corroborated by this. Notably, structural defects appear on the membrane surface when the reaction temperature is increased to 120 °C.
[0107] Therefore, to further control the surface charge of the GO film, the NGO-60 film underwent thermal reduction treatment. Figure 11 As shown in b, the GO film maintains a positive charge throughout the reduction process, reaching its maximum charge density at 140 °C. This is attributed to the thermal decomposition of the oxygen-containing functional groups that provide the negative charge. Figure 10 c). As the reduction temperature further increases, the surface charge density of the GO film gradually decreases. This trend stems from the thermal degradation of positively charged amino groups at high temperatures, which is reflected in the decrease in nitrogen content in the XPS spectrum ( Figure 10 The reduction of the C–N stretching vibration peaks in the FTIR spectrum of the heat-treated film (c, f) and the reduction of the C–N stretching vibration peaks in the FTIR spectrum of the heat-treated film ( Figure 9 b) was verified.
[0108] A noticeable visual change was observed when the membrane transitioned from a negative to a positive charge state. Significant color differences were observed in both PGO and NGO suspensions and in the filtration membrane, with the NGO sample appearing dark brown. SEM images showed that the morphology and thickness of the membrane remained largely unaffected by the change in surface charge state. Figure 12 Raman spectroscopy showed that the ID / IG ratio of the GO film was similar in both positive and negative charge states, indicating that the structural integrity of the GO nanosheets was well maintained during charge transition. When the thermal reduction temperature exceeded 180 °C, the ID / IG ratio exceeded 1.18. Figure 13 (a) indicates that increased structural disorder may weaken the membrane's separation performance. XRD results show that the interlayer spacing gradually decreases with increasing thermal reduction temperature. Figure 13 (b), which is mainly attributed to the removal of surface functional groups.
[0109] Ion separation performance
[0110] The K⁺ / Mg²⁺ separation performance of the charged GO membrane was studied using ion permeation experiments. Figure 13 As shown in d, the ion permeability of the membrane continues to decrease, which is partly attributed to the gradual reduction in interlayer spacing. Figure 13 (b). As the reduction temperature increases, the hydrophilicity of the membrane gradually decreases. Figure 13 The decrease in ion permeability is also due to the reduced hydrophilicity (d). Since ion transport in GO membranes is essentially coupled with water flow, the reduced hydrophilicity weakens the membrane's affinity for water molecules, thus slowing down the co-transport process of water molecules carrying hydrated ions. The K⁺ / Mg²⁺ selectivity shows a trend of first increasing and then decreasing, and is most pronounced at a reduction temperature of 160 °C (T r The maximum value is reached at NGO-160. This optimal condition corresponds to a good balance between surface charge density and interlayer spacing. Figure 11 b and Figure 13 (b). Therefore, the highest selectivity can be attributed to the synergistic effect of size sieving and electrostatic repulsion revealed by molecular dynamics simulations.
[0111] Because of T r The NGO-160 membrane exhibited optimal selectivity while maintaining moderate permeability; therefore, its performance in separating mixed salts was further evaluated under different pH conditions. Zeta potential testing showed that when pH > 3, T... r The surface charge of the NGO-160 membrane changes from positive to negative. Figure 14 (a), and the surface charge density gradually increases with increasing pH. Figure 14 (a). This trend is mainly attributed to the gradual deprotonation of positively charged amino groups into neutral species, and the gradual deprotonation of carboxyl groups to impart a negative charge to the membrane.
[0112] The separation results show that the membrane maintains a certain ion separation capability under both positive and negative charge states. Under the positive charge state, the membrane achieves a K⁺ / Mg²⁺ selectivity of 77, corresponding to a permeability of 0.26 mol m⁻² h⁻¹, which is superior to its performance under the negative charge state. Figure 14 (b). This result is consistent with the conclusions of molecular dynamics simulations, indicating that the electrostatic repulsion between the surface positive charge and K⁺ / Mg²⁺ can effectively modulate its transmembrane transport energy barrier, thereby promoting ion separation. Compared with various ion separation membranes reported in the literature, this positively charged GO membrane exhibits superior overall separation performance (b). Figure 14 (c). Furthermore, the membrane maintained stable performance for 172 hours. Figure 14 The d) indicates its potential for application in actual separation processes.
[0113] References:
[0114] [1] Y. Zhang, S. Zhang, J. Gao, T.-S. Chung. Layer-by-layerconstruction of graphene oxide (GO) framework composite membranes for highlyefficient heavy metal removal. J. Membr. Sci., 515 (2016) 230-237.
[0115] [2] H. Tang, D. Liu, Y. Zhao, X. Yang, J. Lu, F. Cui. Moleculardynamics study of the aggregation process of graphene oxide in water. TheJournal of Physical Chemistry C, 119 (47) (2015) 26712-26718.
[0116] [3] T. Lu, F. Chen. Multiwfn: A multifunctional wavefunctionanalyzer. J. Comput. Chem., 33 (5) (2012) 580-592.
[0117] [4] T. Lu. A comprehensive electron wavefunction analysis toolbox forchemists, Multiwfn. The Journal of chemical physics, 161 (8) (2024) 。
Claims
1. A method for K + / Mg 2+ The separated GO membrane is characterized by, The GO membrane includes a porous support layer and a graphene oxide separation layer disposed on the surface of the porous support layer, wherein the graphene oxide separation layer is an amino-functionalized graphene oxide layer.
2. The method for K according to claim 1 + / Mg 2+ The separated GO membrane is characterized by, The amino-functionalized graphene oxide layer is formed by loading ammonia-modified graphene oxide onto a porous support layer; the graphene oxide separation layer contains amino functional groups corresponding to CN bonds; the GO membrane exhibits positive charge on its surface under acidic conditions.
3. The method for K as described in claim 1 + / Mg 2+ A method for preparing a separated GO membrane, characterized in that, Includes the following steps: (1) Ammonia water was added to the graphene oxide dispersion to react and obtain amino-functionalized graphene oxide; (2) The amino-functionalized graphene oxide is dispersed in water to obtain a dispersion; (3) The dispersion is loaded onto the surface of the porous support layer to form a graphene oxide separation layer; (4) The obtained membrane is thermally reduced to obtain the GO membrane.
4. The preparation method according to claim 3, characterized in that, In step (1), the concentration of the graphene oxide dispersion is 0.1-50 mg / g; the volume ratio of ammonia added to the graphene oxide dispersion is 0.5:(20-100); the concentration range of ammonia is 20-30 wt%; the reaction temperature is 40-120℃, and the reaction time is 2-20 h; the step of collecting the solid product by centrifugation is also included afterward.
5. The preparation method according to claim 3, characterized in that, In step (2), the amount of amino-functionalized graphene oxide added to deionized water is controlled at a mass-to-volume ratio of 0.5 mg:(100-500) mL, and ultrasonic dispersion is performed; and the loading of amino-functionalized graphene oxide on the porous support layer surface is 0.01-0.1 mg / cm³. 2 The heat treatment temperature in step (4) is 120-200℃.
6. The GO membrane according to claim 1 in K + / Mg 2+ Applications in separation.
7. The application according to claim 6, characterized in that, K + Mg preferentially permeates through the GO membrane. 2+ Selective retention; the application is carried out in the pH range of 3 to 11.
8. A method for using the GO membrane of claim 1 to K + / Mg 2+ Molecular dynamics simulation methods for separation performance include the following steps: (1) Establish a nanochannel model of graphene oxide membrane, wherein the graphene oxide membrane is an amino-functionalized graphene oxide membrane or a carboxyl-functionalized graphene oxide membrane. (2) Water-containing chambers are set on both sides of the nanochannel model, and K is set in the feed-side chamber. + Mg 2+ And Cl⁻ used to maintain electroneutrality; (3) Apply external pressure to the nanochannel model and perform non-equilibrium molecular dynamics simulations to obtain K. + and Mg 2+ The transmembrane transport trajectory; (4) Calculate K based on the simulation results + and Mg 2+ The number of ions permeated was determined, and the interaction energy between ions and the graphene oxide membrane and / or the average force potential for ion transport across the membrane were calculated. (5) According to K + With Mg 2+ The differences in the amount of permeate, the differences in the interaction energy between ions and the graphene oxide membrane, and / or the differences in the ion transmembrane transport barrier are used to predict the effect of the charged graphene oxide membrane on K. + / Mg 2+ Separation performance.
9. The simulation method according to claim 8, characterized in that, In step (1), the graphene oxide membrane nanochannel model is constructed based on the Lerf-Klinowski model; the graphene oxide model has dimensions of 25-35 Å in both the x and y directions, and amino or carboxyl functional groups are uniformly distributed at the edges of the graphene oxide sheets; the lateral distance between two graphene sheets is determined by umbrella-shaped sampling simulation, and a 0.8 ohm coefficient is used based on the average force potential curve. A single-layer graphene oxide membrane channel model was constructed at a distance of nm; in step (2), a rigid original graphene layer was set on each side of the nanochannel model as a piston to apply pressure and maintain structural stability; a 90 Å vacuum layer was set in the z direction to eliminate the interaction between the system and its periodic mirror image; the molecular dynamics simulation applied periodic boundary conditions in the three Cartesian directions of x, y and z, and the water molecules were described using the SPC / E model; the molecular dynamics simulation used the Amber force field to describe the interatomic interactions, the non-bonded interatomic interactions were calculated using the Lennard-Jones 12-6 potential and the Coulomb potential, and the cross-interaction parameters were calculated using the Lorentz-Berthelot mixing rule; graphene oxide and graphene sheets All sp² hybridized carbon atoms were considered electrically neutral, and the partial charges of carboxyl and amino groups were determined using the RESP2 method. To maintain the rigidity of the graphene piston, a dihedral torsional barrier was applied. In the nonequilibrium molecular dynamics simulation, the cutoff distance for van der Waals interactions was 0.5-2 nm, and long-range electrostatic interactions were handled using the particle grid Ewald method. The simulation process included energy minimization followed by a 60-70 ns simulation under a canonical ensemble. During the simulation, the carbon sheets of the graphene oxide film were positionally restricted, while the oxygen-containing functional groups remained relaxed. The system temperature was maintained at 295-300 K using a V-rescale thermostat, the initial velocity was randomly generated, the integration time step was 0.5-1 fs, and the trajectory coordinates were recorded every 2-8 ps.
10. The simulation method according to claim 8, characterized in that, The external pressure applied in step (3) is 180-220 MPa, and this external pressure is achieved by applying a force to each carbon atom on the original graphene piston, the force being calculated by the following formula:
11. Among them, f is the force applied to each carbon atom, P is the pressure, A is the piston area, and n is the total number of carbon atoms in the piston. In step (4), the interaction energy between the ions and the graphene oxide film is calculated using the following formula:
12. Among them, Represents the energy of ion-GOGO compounds. and These correspond to the energies of isolated ions and isolated GO nanosheets, respectively. The more negative the value, the stronger the attraction between the ions and the graphene oxide film; The more positive the value, the stronger the repulsion between the ions and the graphene oxide film; In step (4), the energy barrier for ion transmembrane transport is calculated using the average force potential, which is calculated using the adaptive bias force method implemented by the Colvars module in GROMACS. In step (4), the radial distribution function between ions and functional groups is also calculated to characterize the distribution probability of ions near functional groups. The cutoff distance of the radial distribution function is set to half the length of the simulation box. In step (4), the average residence time of ions in the range of 4-6 Å of functional groups is also statistically analyzed to characterize the adsorption and enrichment behavior of ions near functional groups. The ion transmembrane transport trajectory is also analyzed to identify the inter-ion Coulomb repulsion and sequential knock-on transport behavior triggered by the enrichment of ions near carboxyl groups. In step (5), when the amino-functionalized graphene oxide film reacts with Mg... 2+ Showing higher than K + When the transmembrane transport barrier is given, it is predicted that the membrane has K0. + Selective transport performance; when carboxyl-functionalized graphene oxide film is subjected to K + and Mg 2+ When exhibiting similar permeation rates, the membrane is predicted to be effective against K. + / Mg 2+ The selectivity is low; based on the step prediction results, amino-functionalized graphene oxide film was selected as K. + / Mg 2+ Candidate membranes for separation.