Analysis method for alumina membrane filtration under ethanol-water system based on molecular dynamics

Abstract

The invention discloses a molecular dynamics-based alumina membrane filtration analysis method under an ethanol-water system, which is characterized by comprising the following steps: S1, establishing an ethanol water solution model and an alumina membrane model; determining boundary conditions; determining an initial parameter and a potential function; a simulation initialization step; a pressure-driven filtering simulation step; a simulation process trajectory data acquisition step; data analysis and visualization processing are carried out. According to the invention, the distribution and intensity of the hydrogen bond network, the molecular diffusion behavior and the alcohol-water cluster change of the ethanol-water system can be observed at the microscopic scale.

Description

Technical Field

[0001] This invention relates to the field of computational chemistry, and more specifically to an analytical method for alumina membrane filtration in an ethanol-water system based on molecular dynamics. Background Technology

[0002] Wine filtration is an indispensable part of the brewing process, determining not only the clarity and stability of the product but also directly impacting its sensory quality and flavor characteristics. Traditional filtration methods, such as diatomaceous earth filtration and centrifugation, have limitations including complex operation, difficulty in production control, and challenges in ensuring product quality. In recent years, inorganic ceramic membrane separation technology has shown broad application prospects in the food and fermentation industries due to its high-temperature resistance, good chemical stability, and high strength. Alumina ceramic membranes, in particular, are gradually becoming a research hotspot in the field of wine filtration due to their unique physicochemical properties and adjustable pore size distribution. Experimental studies on wine filtration using alumina ceramic membranes have shown that the ceramic membranes effectively remove turbidity from the wine, with minimal difference in physicochemical indicators and taste compared to the unfiltered sample. Furthermore, the filtered wine sample showed no sediment or significant turbidity after being stored at 0°C for 120 hours, demonstrating good stability under room temperature conditions.

[0003] Modern analytical techniques such as infrared spectroscopy, fluorescence spectroscopy, Raman spectroscopy, and nuclear magnetic resonance can be used to analyze the dynamic changes in the hydrogen bond network and cluster structure in the ethanol-water system. For example, three-dimensional fluorescence spectroscopy can capture the differences in ethanol-water clusters in different ethanol-water systems; static laser light scattering can be used to study the molecular cluster structure of ethanol-water colloidal molecules, systematically analyze various fine physicochemical properties of the colloid, explore the existence of irregular hydrogen bond cluster network colloidal structures in the ethanol-water system, and discover that differences in hydrogen bond network and colloidal stability are directly related to the quality and sensory experience of the wine.

[0004] While experimental analysis techniques have yielded a wealth of results regarding the static observation of the microstructure of wine, in-depth research remains lacking on the dynamic mechanisms of changes in the hydrogen bond network and molecular cluster structure during ceramic membrane filtration. In particular, how filtration affects the reconstruction of the hydrogen bond network, the evolution of the molecular cluster structure, and how these changes influence the sensory quality of the wine remain mysteries. This requires a combination of experimental research and molecular simulation methods to reveal the mechanisms by which ceramic membrane filtration affects the structure of wine at the molecular level.

[0005] To overcome the aforementioned deficiencies in existing technologies, this invention employs molecular dynamics simulations to explain, from a microscopic perspective, the dynamic changes in the hydrogen bond network of the ethanol-water system during alumina membrane filtration and the changes in the number of clusters of different sizes before and after filtration. First, models with different alcohol concentrations and membrane pore sizes are constructed. After structural optimization, simulations are run under different temperatures and pressures to obtain molecular trajectory files under different variable controls. Finally, the trajectories are analyzed, data is extracted, and results including changes in hydrogen bond network, mean square displacement, diffusion coefficient, and cluster proportion are generated. The resulting data are used to optimize and update the original model, ultimately yielding microscopic observations of the ethanol-water system after nanopore-confined filtration. Summary of the Invention

[0006] In order to overcome the above-mentioned defects of the prior art, the purpose of this invention is to provide an analytical method for alumina membrane filtration in an ethanol-water system based on molecular dynamics.

[0007] Unlike existing technologies, this invention enables the observation of the distribution and strength of hydrogen bond networks, molecular diffusion behavior, and changes in alcohol-water clusters in the ethanol-water system at the microscopic scale.

[0008] To achieve the objectives of this invention, the technical solution adopted is as follows: An analytical method for alumina membrane filtration in an ethanol-water system based on molecular dynamics includes the following steps: S1, Steps for establishing the ethanol-water solution model and the alumina film model: Establish ethanol-water solution models and alumina film models in a simulation box; S2, Steps to determine boundary conditions: The boundary conditions are determined by using periodic boundary conditions over the entire simulation box; S3, Steps to determine initial parameters and potential function: The force fields used by various molecules are determined, and nonbonded interactions are calculated using the LJ potential function and Coulomb's law. S4, Simulation initialization steps: The model is subjected to energy minimization and the system is pre-equilibrated using the NVT ensemble; S5, Pressure-Driven Filter Simulation Steps: External pressure was applied to simulate the filtration of solutions through an alumina membrane under different pressures; S6, Simulation process trajectory data acquisition steps: The system records the coordinates, velocity, force, and potential energy of each particle during the simulation at regular intervals to obtain trajectory data during the simulation. S7, Data Analysis and Visualization Steps: The trajectory files and calculation results were analyzed, and the analysis results, including changes in hydrogen bond network, mean square displacement, diffusion coefficient, and cluster proportion, were extracted and then visualized.

[0009] In a preferred embodiment of the present invention, step S1, establishing the ethanol-water solution model and the alumina film model, specifically comprises: The first step is to calculate the stoichiometric ratio of ethanol to water molecules in the solution based on the target ethanol concentration, determine the specific number of each molecule in the simulation box, and verify the rationality of the force field parameters and model construction by comparing the consistency between the simulation system and the experimentally measured density. The second step involves using Materials Studio software to construct an α-Al2O3 crystal model. Nanopores with square cross-sections and pore sizes of 2 nm, 2.5 nm, and 3 nm are cut along specific crystal planes. The inner surface of the nanopores is then modified with hydroxylation to simulate the hydroxyl terminal structure formed on the surface of alumina in an aqueous environment, thus more realistically reflecting the surface chemical state under actual conditions. The third step involves placing the hydroxylated alumina membrane obtained in the second step in the center of the box in the final simulation system, and applying positional constraints to its atoms to maintain rigidity, thereby ensuring that it remains structurally stable and maintains effective molecular sieving function during the filtration process.

[0010] In a preferred embodiment of the present invention, step S2, determining the boundary conditions, specifically involves applying periodic boundary conditions to the entire simulation box in the x, y, and z directions.

[0011] This approach aims to reduce computational costs and eliminate size effects.

[0012] In a preferred embodiment of the present invention, step S3, determining the initial parameters and potential function, specifically comprises: Water molecules were modeled using the OPC3 water model, ethanol molecules using the GAFF force field, and alumina crystals using the ClayFF force field of metal oxides. The molecular dynamics simulation of the large system was run in the GROMACS software package, and nonbonded interactions were calculated using the LJ potential function and Coulomb's law.

[0013] In a preferred embodiment of the present invention, step S4 simulates the initialization step: The main parameter settings for energy minimization of the model include: using the STEEP steepest descent method, setting the force convergence to 500 kJ / mol / nm, selecting the Verlet nearest neighbor list algorithm for truncation, and using the Cut-off and PME methods for van der Waals and Coulomb interactions, respectively, with a cutoff radius of 1.2 m. A Freeze setting was used to freeze the graphene plate and filter membrane to prevent large deformation and displacement. The system was then pre-equilibriumed using the NVT ensemble to achieve thermodynamic equilibrium. During the pre-equilibrium stage, no pressure was applied to the graphene plate, and the alumina membrane and graphene were frozen to allow the liquid phase to fully equilibrium before observing the solution density.

[0014] This step ensures that the system can run stably without crashing during the simulation, and compares the solution density in the model with the density measured experimentally to verify the accuracy of the model.

[0015] In a preferred embodiment of the present invention, step S5, pressure-driven filtration simulation step: By setting the centroid traction module in GROMACS software, the xy direction of the graphene plate is frozen, allowing free movement in the z direction. Pressure is applied in the z direction to the graphene plate, causing it to push forward and squeeze the solution through the alumina filter membrane, thus simulating the effect of pressure during the actual filtration process.

[0016] This step uses an unbalanced simulation method, where the applied pressure is usually much greater than the actual pressure, in order to improve the signal-to-noise ratio within a nanosecond-level simulation time.

[0017] In a preferred embodiment of the present invention, step S6, the simulation process trajectory data acquisition step, involves outputting and acquiring the coordinates, energy, and force information of the atoms in step S5 every 10 ps, ​​and combining this with the coordinates of hydrogen and oxygen atoms required for calculating hydrogen bonds in the system to obtain trajectory data during the simulation process, and studying the dynamic changes of the hydrogen bond network.

[0018] In a preferred embodiment of the present invention, the visualization process in S7 is to use VMD software to visualize the simulated trajectory, which can clearly show the solution movement and dynamic changes of the hydrogen bond network during the entire simulation process from running to equilibrium. The diffusion behavior was studied using the GROMACS data analysis script, which is a built-in MSD calculation program in the software. The command is gmx - gmx msd -f prod.trr -s prod.tpr -o msd.xvg to calculate the mean square displacement (MSD) of the molecules, and the diffusion coefficient of different molecules in the simulation process is obtained by fitting the Einstein formula. The changes in clusters in the system are analyzed by data judgment and analysis to determine whether they are in the same cluster by analyzing the distance between oxygen atoms in different molecules, thus obtaining clusters of different sizes.

[0019] The specific data analysis method involves inputting a simulated trajectory coordinate file and analyzing the distances between the coordinates of each atom. If the distance between two oxygen atoms is between 2.5 and 2.9 Å, it is determined that a hydrogen bond has formed. The residue names of the two molecules are then stored for display in subsequent visualization software. This can result in large clusters of multiple molecules connected together, or it can produce free ethanol molecules. Therefore, the influence of various conditions on the clusters can be determined by statistically analyzing the changes in the proportion of each type of cluster.

[0020] The analysis results were plotted using Origin software to analyze the changing trends of various parameters before and after solution filtration.

[0021] The beneficial effects of this invention are as follows: By employing computer simulations and molecular dynamics research methods, we can reveal the differences between different microscopic scales. The changes in the hydrogen bond network and cluster quantity of the ethanol-water system after passing through an alumina ceramic membrane can compensate for the lack of experimental explanation of the mechanism of these changes. Attached Figure Description

[0022] Figure 1 This is a flowchart of the present invention.

[0023] Figure 2 is a schematic diagram of the filtering model of the present invention. Figure 2a Schematic diagram of model density verification Figure 2b .

[0024] Figure 3 is a schematic diagram comparing the calculated and experimental values ​​of water and ethanol molecule permeation rates and water molecule permeation flux under different pressures according to the present invention. Figure 3a and Figure 3b .

[0025] Figure 4 This is a schematic diagram comparing the changes in hydrogen bond lifetime under different temperatures and pressures according to the present invention.

[0026] Figure 5 This is a schematic diagram comparing the changes in the proportion of clusters in the system at different temperatures according to the present invention.

[0027] Figure 6 This is a schematic diagram comparing the changes in the proportion of clusters in the system under different pressures according to the present invention.

[0028] Figure 7 This is a schematic diagram comparing the changes in hydrogen bond lifetime under different pore sizes according to the present invention.

[0029] Figure 8This is a schematic diagram comparing the changes in the proportion of clusters in the system with different pore sizes according to the present invention. Detailed Implementation

[0030] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to figures and embodiments. However, it should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of the invention. Furthermore, in the following descriptions, well-known structures and technologies are omitted to avoid unnecessarily obscuring the concept of the invention.

[0031] like Figure 1 An analytical method for alumina membrane filtration in an ethanol-water system based on molecular dynamics is shown, comprising the following steps: S1, Steps for establishing the ethanol-water solution model and the alumina film model: The specific steps for establishing the ethanol-water solution model and the alumina film model are as follows: The first step is to calculate the stoichiometric ratio of ethanol to water molecules in the solution based on the target ethanol concentration, determine the specific number of each molecule in the simulation box, and verify the rationality of the force field parameters and model construction by comparing the consistency between the simulation system and the experimentally measured density.

[0032] The specific operation in this embodiment is as follows: Ethanol and water molecules were drawn using Material Studio software, and the structure was optimized using its built-in CASTEP module. Then, the Amorphous Cell module was used to generate solution boxes with different numbers of ethanol and water molecules according to different alcohol concentrations, namely Et10, Et20, Et30, Et40, and Et50, with a side length of 50 Å.

[0033] The second step involves using Materials Studio software to construct an α-Al2O3 crystal model. Nanopores with square cross-sections and pore sizes of 2 nm, 2.5 nm, and 3 nm are cut along specific crystal planes. The inner surface of the nanopores is then modified with hydroxylation to simulate the hydroxyl terminal structure formed on the surface of alumina in an aqueous environment, thus more realistically reflecting the surface chemical state under actual conditions.

[0034] The specific operation in this embodiment is as follows: The alumina cell was selected as α-Al2O3, and its cell file was exported from a crystal library website. Then, Materialstudio software was used to cut its (0001) face and expand it into a supercell. Holes were drilled in the center of the film to form square holes with side lengths of 2nm, 2.5nm, and 3nm. The film wall was hydroxylated to simulate the real environment of alumina in solution.

[0035] The third step involves placing the hydroxylated alumina membrane obtained in the second step in the center of the box in the final simulation system, and applying positional constraints to its atoms to maintain rigidity, thereby ensuring that it remains structurally stable and maintains effective molecular sieving function during the filtration process.

[0036] The specific operation in this embodiment is as follows: The solution, alumina membrane, and graphene plate are assembled to form a complete solution filtration model, as shown in Figure 2.

[0037] S2, Steps to determine boundary conditions: Periodic boundary conditions are applied to the entire simulation box in the x, y, and z directions. This is done to reduce computational cost and eliminate size effects. A freeze setting is used to fix the alumina film in the middle to prevent lateral drift of the film due to pressure.

[0038] S3, Steps to determine initial parameters and potential function: The force fields used for various molecules were determined, and nonbonded interactions were calculated using the LJ potential function and Coulomb's law. The OPC3 water model was used for water molecules, the GAFF force field was used for ethanol molecules, and the ClayFF force field of metal oxides was used for alumina crystals. The molecular dynamics simulation of the large system was run in the GROMACS software package, and nonbonded interactions were calculated using the LJ potential function and Coulomb's law.

[0039] In this embodiment, the alumina film uses the ClayFF force field commonly used for metal oxides, the water uses the OPC3 water model (which better reflects the hydrogen bonding characteristics in the reaction system), and the GAFF force field is used for ethanol and other small organic molecules. After energy minimization convergence, the long-range electrostatic interactions and van der Waals interactions of the equilibrium phase and the production phase are calculated using PME and Cut-off methods, respectively, with a cutoff distance of 12 Å.

[0040] S4, Simulation initialization steps: The model underwent energy minimization, with key parameter settings including: using the STEEP steepest descent method, setting the force convergence to 500 kJ / mol / nm, selecting the Verlet nearest neighbor algorithm for truncation, and using the Cut-off and PME methods for van der Waals and Coulomb interactions, respectively. In this embodiment, the truncation radius was 1.2 nm for both methods. A Freeze setting was used to freeze the graphene plate and filter membrane to prevent significant deformation and displacement. The system was then pre-equilibrated using the NVT ensemble to achieve thermodynamic equilibrium. During the pre-equilibrium stage, no pressure was applied to the graphene plate; instead, the alumina film and graphene were frozen to allow the liquid phase to fully balance before observing the final liquid phase density. The system underwent pre-equilibrium treatment to achieve thermodynamic equilibrium.

[0041] This step ensures that the system can run stably without crashing during the simulation, and compares the solution density in the model with the density measured experimentally to verify the accuracy of the model.

[0042] In this embodiment, the simulations were all conducted in the NVT system. First, the graphene and alumina films were frozen and pre-equilibrated for 2 ns. After equilibrium, the density on the ethanol-water solution side was measured to be close to the experimental value, as shown in Figure 2.

[0043] S5, Pressure-Driven Filter Simulation Steps: External pressure was applied to simulate the filtration of solutions through an alumina membrane under different pressures. The main filtration conditions were determined by the permeation rates of water and ethanol molecules and the permeation flux of water molecules. The graphene plate was frozen in the xy direction using the centroid traction module of the GROMACS software, while the z direction was allowed to move freely. Pressure was applied in the z direction to the graphene plate to push it forward and squeeze the solution through the alumina filter membrane, thus simulating the effect of pressure during the actual filtration process.

[0044] This step uses an unbalanced simulation method, where the applied pressure is usually much greater than the actual pressure, in order to improve the signal-to-noise ratio within a nanosecond-level simulation time.

[0045] In this embodiment, the center-of-mass traction setup of GROMACS is used to apply a pressure of 700 MPa to the graphene plate, causing the movable rigid wall to move forward, thereby simulating a pressure-driven filtration process. The pressure applied during the simulation is much greater than the experimental value; this is a non-equilibrium simulation. NEMD This is a common method used to reduce thermal noise and improve the signal-to-noise ratio, thereby enabling the simulation to obtain statistically meaningful results within a reasonable time. The generated phase is simulated at 298 K with a runtime of 10 ns. Sampling is performed on the last 4 ns to avoid unreasonable results caused by the violent perturbation of the hydrogen bond network within the system by the solution-side motion in the first half of the simulation.

[0046] The pressure settings for the nonequilibrium simulation have been explained above. The reason for setting the simulated pressure higher than the actual pressure is to obtain statistically reliable results within nanosecond timeframes, i.e., to reduce the influence of thermal noise in the simulation and thus improve the signal-to-noise ratio. However, it is still impossible to correlate the results with the pressure in the actual experiment. Therefore, we chose to verify the results from the perspective of the pressure-driven effect on membrane permeability changes (see Figure 3), which is also a commonly used method in some nonequilibrium simulations.

[0047] Flux is defined as the number of molecules that permeate through the membrane per unit time, i.e.: Flux = Ns / t; Where Ns is the number of molecules passing through in time t, and t represents the selected simulation time period, the permeation rate under different pressures is statistically analyzed and linearly fitted. The permeation rate under low pressure is obtained by extrapolating the fitted line forward (the fitting formula is shown in the figure). The resulting flux can be further correlated with the geometric parameters of the model to obtain the model's permeability, denoted by Pe (unit: kg / (m³)). 2 The formula for calculating h·bar is: Pe=NsM / N A At Where Ns represents the number of molecules passing through, M represents the relative molecular weight, and N A denoted by Avogadro's constant, A represents the effective area of ​​the filtration membrane, and t represents the selected simulation time period. Since the simulation systems used were all low-ethanol concentration systems, water permeation was the dominant process, and the influence of ethanol molecule permeation behavior could be ignored. Comparing the water permeation flux in the simulated system with the experimental permeation flux of deionized water filtered by an alumina ceramic membrane, the error was found to be around 5%, indicating that the model can accurately reflect the influence of pressure-driven water molecule permeation behavior.

[0048] In this embodiment, the simulation of the large system used in steps S2, S3, S4, and S5 is all completed within the GROMACS software package.

[0049] S6, Simulation process trajectory data acquisition steps: The system records the coordinates, velocity, force, and potential energy of each particle during the simulation at regular intervals to obtain trajectory data during the simulation.

[0050] The simulation process trajectory data acquisition step involves outputting and acquiring the coordinates, energy, and force information of the atoms in step S5 every 10 ps. The trajectory data during the simulation process is obtained by combining the coordinates of hydrogen and oxygen atoms required for calculating hydrogen bonds in the system, and the dynamic changes of the hydrogen bond network are studied.

[0051] The main focus is on the results of studying the changes in hydrogen bond lifetime and cluster number under different conditions. S7, Data Analysis and Visualization Steps: The trajectory files and calculation results were analyzed, and the analysis results, including changes in hydrogen bond network, mean square displacement, diffusion coefficient, and cluster proportion, were extracted and then visualized.

[0052] The visualization process in S7 uses VMD software to visualize the simulated trajectory, which clearly shows the solution movement and dynamic changes of the hydrogen bond network throughout the simulation from start to finish. The diffusion behavior was studied using the GROMACS data analysis script, which is a built-in MSD calculation program in the software. The command is gmx - gmx msd -f prod.trr -s prod.tpr -o msd.xvg to calculate the mean square displacement (MSD) of the molecules, and then the Einstein formula was used to fit the data to obtain the diffusion coefficient of different molecules in the simulation process. The changes in clusters within the system are analyzed using data-driven judgment to determine whether oxygen atoms in different molecules belong to the same cluster, thus obtaining clusters of different sizes. Specifically, the simulation-generated trajectory coordinate file is input, and the distances between the coordinates of each atom are analyzed. If the distance between two oxygen atoms is between 2.5 and 2.9 Å, it is determined that a hydrogen bond has formed. The residue names of the two molecules are then stored for later visualization in visualization software. This can result in large clusters of multiple molecules connected together, as well as free ethanol molecules. Therefore, the influence of various conditions on clusters can be determined by statistically analyzing changes in the proportion of different cluster types.

[0053] The analysis results were plotted using Origin software to analyze the changing trends of various parameters before and after solution filtration.

[0054] Among them, the average lifetime of hydrogen bonds τ Averaging all autocorrelation functions of all hydrogen bond existence functions (0 or 1), and considering the hydrogen bonds at time t. , The integral gives the average lifetime of hydrogen bonds. τ The estimate is given by the following formula:

[0055] It represents the average duration of hydrogen bonds, and therefore can accurately describe the stability of the hydrogen bond network structure in the ethanol-water system.

[0056] Depend on Figure 4 As can be seen, the increase in temperature reduces the hydrogen bond lifetime of the system, which means that the increase in temperature will destroy the stability of the hydrogen bond network in the alcohol-water system. The explanation for this is that the increase in temperature will lead to the intensification of molecular thermal motion in the system, which will destroy the formed hydrogen bond structure and cause part of the hydrogen bond network to collapse.

[0057] As the pressure applied to the graphene plate on the left increases, the hydrogen bond lifetime initially rises and then falls, peaking between 500 MPa and 700 MPa with minimal variation. Changes in pressure not only affect the rate at which molecules pass through the alumina film but also compress the alcohol-water system. At lower pressures, the piston-like compression effect reduces the distance between alcohol-water molecules, making the system more compact and the hydrogen bonds more stable, resulting in a longer hydrogen bond lifetime. However, when the pressure exceeds a critical value, the existing hydrogen bonds break, and the hydrogen bond network structure is disrupted.

[0058] Clusters are an important part of the formation of hydrogen bond network structure in alcohol-water systems. The size, structure and composition of clusters have a very important impact on the taste of alcohol. The inventors have performed clustering treatment on ethanol and water molecules.

[0059] Using the atomic coordinates in the trajectory file, if the distance between the oxygen atoms of two molecules is between 2.7 and 2.9 angstroms, then hydrogen bonds are considered to have formed, and they are classified into the same cluster. The result is 3-8 molecule clusters with different proportions.

[0060] This statistical method counts the various combinations of ethanol and water molecules in the system, rather than clusters containing a specific number of ethanol or water molecules.

[0061] See Figure 5 , 6 Statistical results show that large molecular clusters, i.e., octet clusters, dominate the clusters in the system, while small molecular clusters (such as 3- and 4-molecule clusters) account for only about a quarter of the total number of clusters.

[0062] exist Figure 5 , 6 The first column shows the initial state of the system before filtration through the ceramic membrane. As can be seen from the figure, the changes caused by increasing temperature vary across different ethanol concentration regions. In the low concentration region, as shown in figures a and b (representing systems at Et10 and Et20 concentrations respectively), increasing temperature leads to a decrease in the proportion of 8-molecule clusters and an increase in the proportion of small molecule clusters, but 8-molecule clusters still dominate the system. In the high concentration region, as shown in figures ce (representing systems at Et30, Et40, and Et50 concentrations respectively), increasing temperature increases the proportion of 8-molecule clusters and decreases the proportion of small molecule clusters.

[0063] observe Figure 7 The trend showed that the effect of increased pressure on clusters was uniform. A horizontal comparison of results at different concentrations did not yield any concentration-specific conclusions. After increasing the pressure, the proportion of 8-molecule clusters decreased, while the proportion of small molecule clusters increased.

[0064] Through extensive research, the inventors discovered that: The structure and size of the ceramic membrane pores also play a role in the filtration effect, and setting different pore sizes for evaluation is an important aspect.

[0065] Therefore, the inventors used Materials Studio software to create holes in the alumina film, setting the side lengths of the square holes to 2×2, 2.5×2.5, and 3×3 nm. The same simulation method was used to statistically analyze the system under different pore sizes. To avoid errors caused by temperature and pressure, the temperature at which the phase was generated was set to room temperature, and the pressure to 700 MPa.

[0066] See Figure 7 , 8 As the pore size increases, the hydrogen bond lifetime gradually decreases, meaning the stability of the hydrogen bond network gradually weakens. As shown in the figure, the increase in pore size has little effect on clusters; there is a slight tendency for 8-molecule clusters to increase. This can be reasonably explained by considering the effect of membrane pores on clusters in the previous section. Increased pore size leads to a weakening of the barrier effect on large molecular clusters; only a small portion of 8-molecule clusters are destroyed, while most retain their original structure and pass through the membrane pores to the other side. Therefore, increasing the pore size reduces the filtration effect, resulting in filtered wine that is closer to its original state and has less impact on the taste. The above illustrates and describes the basic principles, main features, and advantages of the invention.

[0067] Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the present invention. Various changes and modifications can be made to the present invention without departing from its spirit and scope. All such changes and modifications fall within the scope of the present invention as claimed, which is defined by the appended claims and their equivalents.

Claims

1. An analysis method of an alumina membrane filtration under an ethanol-water system based on molecular dynamics, characterized by, Comprising the following steps: S1, establishing an ethanol aqueous solution model and an alumina membrane model step: Establishing an ethanol aqueous solution model and an alumina membrane model in a simulation box; S2, determining boundary conditions step: Determine the boundary conditions by applying periodic boundary conditions to the entire simulation box; S3, determining initial parameters and potential function step: Determine the force field adopted by various molecules, and calculate non-bonding interactions using LJ potential function and Coulomb's law; S4, simulation initialization step: Perform energy minimization on the model and pre-equilibrate the system through NVT ensemble; S5, pressure-driven filtration simulation step: Apply external pressure to simulate the filtration of the solution on the alumina membrane under different pressures; S6, simulation process trajectory data collection step: Record the coordinates, velocities, forces and system potential energy data of each particle in the simulation process to obtain the trajectory data in the simulation process; S7, data analysis and visualization processing step: Analyze the trajectory file and calculate the results, extract the analysis results including hydrogen bond network, mean square displacement, diffusion coefficient and cluster proportion change, and perform visualization processing.

2. A method of analyzing an alumina membrane filtration in an ethanol-water system based on molecular dynamics according to claim 1, characterized in that, Said S1, establishing an ethanol aqueous solution model and an alumina membrane model step specifically: First, calculate the stoichiometric ratio of ethanol and water molecules in the solution according to the target ethanol concentration, determine the specific number of each molecule in the simulation box, and verify the rationality of the force field parameters and model construction by comparing the consistency of the simulation system with the experimental measurement density; Second, use Materials Studio software to construct an alpha-Al2O3 crystal model, cut it along a specific crystal plane to obtain a nanochannel with a square cross-section and a pore size of 2 nm, 2.5 nm and 3 nm, respectively, and then modify the inner surface of the nanochannel with hydroxyl groups to simulate the hydroxyl terminal structure formed on the alumina surface in the aqueous environment, and more realistically reflect the surface chemical state under actual conditions; Third, in the final simulation system, place the hydroxylated alumina membrane obtained in the second step in the center of the box, and apply position constraints to its atoms to maintain rigidity.

3. A method for analyzing an alumina membrane filtration in an ethanol-water system based on molecular dynamics as claimed in claim 1, wherein, Said S2, determining boundary conditions step specifically applies periodic boundary conditions to the entire simulation box in the xyz direction.

4. A method of analyzing alumina membrane filtration in an ethanol-water system based on molecular dynamics as claimed in claim 1, wherein, Said S3, determining initial parameters and potential function step specifically: Water molecules use OPC3 water model, ethanol molecules use GAFF force field, alumina crystals use ClayFF force field for metal oxides, and molecular dynamics simulation of large systems is run in GROMACS software package, using LJ potential function and Coulomb's law to calculate non-bonding interactions.

5. A method for analyzing an alumina membrane filtration in an ethanol-water system based on molecular dynamics according to claim 1, characterized in that, Said S4, simulation initialization step: Perform energy minimization on the model, and pre-equilibrate the system through NVT ensemble to achieve thermodynamic equilibrium state of the system, pre-equilibrate the system without applying pressure to the graphene plate, freeze the alumina membrane and graphene, and allow the liquid phase to fully equilibrate to finally observe the liquid phase density.

6. The method of claim 1, wherein the S5, pressure-driven filtration simulation step: ​ Through the setting of the center of mass traction module of the GROMACS software, the xy direction of the graphene plate is frozen, the z direction is free to move, and the z direction pressure is applied on the graphene plate to make it push forward to extrude the solution through the alumina filter membrane to simulate the influence of pressure in the actual filtration process.

7. The method of claim 1, wherein the simulation of the filtration of the ethanol-water system through the alumina membrane is based on molecular dynamics, and the method further comprises: S6, collecting trajectory data during the simulation, wherein the trajectory data is obtained by outputting the coordinates of the atoms, the energy, and the force information of the atoms in the step S5 every 10 ps, and collecting the hydrogen atom and oxygen atom coordinates required for calculating the hydrogen bonds in the system, and researching the dynamic changes of the hydrogen bond network.

8. The method of claim 1, wherein the simulation of the filtration of the ethanol-water system through the alumina membrane is based on molecular dynamics, and the method further comprises: S7, visualizing the trajectory data during the simulation, wherein the trajectory data is visualized by using the VMD software, and the movement of the solution and the dynamic changes of the hydrogen bond network during the simulation from the running to the equilibrium process can be clearly seen. The diffusion behavior of the solution is researched by using the GROMACS data analysis script (the MSD calculation program provided by the software, the instruction is gmx - gmx msd -f prod.trr -s prod.tpr -o msd.xvg) to calculate the mean square displacement (MSD) of the molecules, and the diffusion coefficients of the different molecules in the simulation process are obtained by fitting the Einstein formula. The changes of the clusters in the system are analyzed by judging the distances between the oxygen atoms in different molecules to determine whether they are in the same cluster, and then the clusters with different sizes are obtained. The analysis results are processed by using the Origin software to draw graphs, and the trends of the changes of various parameters before and after the filtration of the solution are analyzed. ​ ​

Images