Methods, equipment and media for simulating acidic site placement in molecular sieves

Abstract

This invention relates to a method, apparatus, and medium for simulating the placement of acidic sites in molecular sieves. The method includes: S1, generating all-silica molecular sieve nanoparticles; S2, randomly inserting a template molecule into a void, selecting the nearest silicon atom for isomorphic substitution of aluminum atoms, as the initial guessed structure; S3, obtaining the relaxed structure of the adsorbed state based on machine learning potential, and calculating the chemical potential change; S4, deciding whether to accept the current structure according to the Metropolis criterion; S5, if the cumulative number of consecutive rejections reaches a set threshold, ending the simulation and proceeding to S6; S6, after determining the aluminum placement, randomly removing a template molecule, randomly adding hydrogen to any bridging oxygen connected to the nearest aluminum framework, obtaining the relaxed structure and changing the position of the hydrogen, determining the placement of acidic sites based on the lowest energy, and repeating this process until all acidic sites are placed; S7, calculating and outputting physical property parameters. Compared with the prior art, this invention has the advantages of short processing time, high accuracy, and strong interpretability.

Description

Technical Field

[0001] This invention relates to the field of molecular sieve technology, and in particular to a method, equipment and medium for simulating the acidic site placement of molecular sieves. Background Technology

[0002] The widespread application of molecular sieves in heterogeneous catalysis stems from their unique pore structure, tunable acidic sites, and excellent shape-selective catalytic performance, making them indispensable catalytic materials in petrochemical, energy conversion, and environmental protection fields. Therefore, modeling the molecular sieve structure and simulating it using theoretical chemistry methods will help to understand its catalytic performance, thus providing a strong basis for catalyst screening and rational design.

[0003] Existing simulation methods can accurately indicate the tendency of acidic sites to fall into steady-state, bulk structures, but they are difficult to explain the falling-in characteristics of nanostructures during growth. The latter has been shown to have excellent catalytic activity in many cases, but its interpretability is not high.

[0004] Furthermore, for molecular sieve nanostructures of a certain size, the potential acidic sites are numerous and cannot be directly simulated using enumeration methods. Meanwhile, the distribution of acidic sites on the surface and in the bulk region is of greater interest than whether a specific location can be occupied. Monte Carlo (MC) simulations estimate the average properties of the system through random sampling: during the simulation, a series of states are continuously generated, and the evolution of the configuration is determined by pre-defined rules. Grand Canonical Monte Carlo (GCMC) simulations are Monte Carlo simulations performed on a grand canonical ensemble, allowing the system to exchange particles with its surroundings, and determining the evolution of the configuration through changes in chemical potential before and after the reaction.

[0005] In the fields of materials design and catalytic reaction simulation, traditional methods rely on quantum mechanical calculations, such as density functional theory (DFT) or empirical potential functions, to construct potential energy surfaces to drive molecular dynamics (MD) or Monte Carlo simulations. However, in the study of the thermodynamic properties of open systems, existing techniques still face significant bottlenecks in balancing computational efficiency and accuracy, as well as adaptability to complex systems: while first-principles calculations can provide high-precision potential energy surfaces, they consume enormous computational resources and are difficult to handle large systems or long-term simulations; while molecular force fields or empirical potential functions (such as Lennard-Jones potential and ReaxFF potential) are computationally efficient, their parameterization process depends on specific systems and cannot accurately describe complex interactions such as chemical bond formation / breaking and charge transfer, especially when dealing with multi-component catalytic surface reactions, where the simplification assumptions introduced can lead to serious deviations in adsorption energy and transition state sites. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of existing simulation methods in terms of accuracy and efficiency, and to provide a method, equipment and medium for simulating the acidic site placement of molecular sieves with higher accuracy, higher efficiency and stronger interpretability.

[0007] The objective of this invention can be achieved through the following technical solutions:

[0008] According to a first aspect of the present invention, a method for simulating the acidic site placement of molecular sieves is provided, comprising:

[0009] S1. Generate all-silica molecular sieve nanoparticles with a set topology and size;

[0010] S2. Randomly insert a template molecule into the gap and select the nearest silicon atom to perform aluminum atom isomorphic substitution, as the initial guess structure of the molecular sieve particle-template adsorption state.

[0011] S3. Based on machine learning potential, obtain the relaxed structure of the adsorbed state and calculate the chemical potential change of the adsorption reaction;

[0012] S4. Based on the chemical potential change obtained in S3, decide whether to accept the current structure according to the Metropolis criterion;

[0013] S5. Update the current structure and go to S2 to generate a new initial guess structure based on the current structure. Repeat S2 to S5 and record the number of times the relaxed configuration is rejected consecutively. If the current number of consecutive rejections reaches the set threshold, end the simulation and go to S6.

[0014] S6. After determining the aluminum position, randomly remove a template agent molecule and randomly add hydrogen to any bridging oxygen connected to the framework aluminum closest to the original position of the template agent molecule. Based on the machine learning potential, obtain the relaxed structure and change the position of the hydrogen to connect with different bridging oxygens in the framework aluminum. Based on the machine learning potential, obtain the relaxed structure respectively. Determine the acid site position based on the lowest energy. Repeat S6 until all acid sites are in place.

[0015] S7. Based on the final molecular sieve-template structure, calculate and output the physical property parameters.

[0016] Preferably, the voids in S2 include the voids on the outer surface and the voids in the molecular sieve channels.

[0017] Preferably, in S3, the architecture of the machine learning potential specifically includes:

[0018] 1) Machine learning can read the original structure of a species and the types and three-dimensional spatial coordinates of each atom;

[0019] 2) Encode the environment surrounding each atom using a centrosymmetric function to obtain a high-dimensional feature vector, where the high-dimensional feature vector corresponding to the i-th atom is represented as: ;

[0020] 3) Each high-dimensional feature vector is mapped to a partial energy value under the action of the same fully connected layer. The total energy of the system is equal to the sum of all the partial energies, expressed as:

[0021]

[0022] in, It is a fully connected function, which is a mapping from a high-dimensional eigenvector space to the real number field.

[0023] Preferably, the chemical potential change in S3 is characterized by the Gibbs free energy of the reaction, expressed as:

[0024]

[0025]

[0026] The corresponding chemical reaction is:

[0027]

[0028] In the formula: To introduce the first n The chemical potential change of each template molecule; To introduce the first Chemical potential following a template molecule To introduce the first Chemical potential in front of a template molecule Represents the electronic energy term obtained based on the machine learning potential; C is a constant when the template agent and reaction temperature are selected, where: For newly inserted template Zero-point energy correction term; For the newly inserted template agent isobaric heat capacity integral correction term, For isobaric heat capacity, The reaction temperature, For newly inserted template Entropy correction term; The standard Gibbs free energy of formation for each species in a given state of matter and temperature in a corresponding chemical reaction; It is a hydroxide ion. Silicon, It is aluminum. Indicates the state of matter in an aqueous solution. It indicates the state of a solid substance.

[0029] Preferably, in step S4, the decision on whether to accept the current structure is made according to the Metropolis criterion, specifically as follows:

[0030] 1) Calculate the probability of accepting the current structure, expressed as:

[0031]

[0032] In the formula: To introduce the first n The chemical potential change of each template molecule, Boltzmann's constant, The reaction temperature;

[0033] 2) Generate a random number in the range [0,1], and the probability... Perform a size comparison; if the random number is smaller than... If the current structure is accepted, then the current structure is accepted; otherwise, the current structure is rejected.

[0034] Preferably, in step S7, the outermost layer is set as the surface-bulk interface within angstroms, the number of silicon atoms and aluminum atoms in the surface and bulk phases are counted, and physical property parameters are calculated; the physical property parameters include chemical composition, total silicon-aluminum ratio, bulk silicon-aluminum ratio, surface silicon-aluminum ratio, surface aluminum, and bulk aluminum.

[0035] Preferably, step S6 specifically includes the following sub-steps:

[0036] 1) After determining the location of aluminum, a template agent molecule is randomly removed, and hydrogen is added to all the bridging oxygens connected to the framework aluminum closest to the original position of the template agent molecule to generate multiple candidate sites for Brønsted acid.

[0037] 2) Based on machine learning potential, obtain the relaxed configuration and energy of all candidate points;

[0038] 3) Select the configuration with the lowest energy after relaxation as the current structure and update it, then return to 1) and continue the cycle until all template agent molecules are removed.

[0039] Preferably, step S7, which calculates and outputs physical property parameters based on the final molecular sieve-template agent structure, specifically includes the following sub-steps:

[0040] S701, Identify aluminum in the surface and bulk phases by partition:

[0041] 1) Calculate and obtain the centroid of the molecular sieve nanoparticles;

[0042] 2) Take the maximum distance from the bulk atom to the center of mass as the radius of the cluster, and set the angstrom within the cluster radius as the surface-bulk interface, and count the number of silicon atoms and aluminum atoms on the surface and in the bulk.

[0043] S702. Identify aluminum in sections of straight channels and straight-channel-sinusoidal-channel intersections:

[0044] 1) Perform Voronoi partitioning on the cluster;

[0045] 2) Take the maximum distance from the bulk atom to the centroid as the radius of the cluster, and set the angstrom within the cluster radius as the surface-bulk interface. Obtain all Voronoi nodes in the bulk phase, as well as the distance of each node to the nearest atom.

[0046] 3) Obtain the maximum value of the distance from the Voronoi node to the nearest atom, with the maximum value as the upper bound and c1 times the maximum value as the lower bound, and retain all nodes whose distance to the nearest atom is between the upper and lower bounds; where c1 is a set constant between 0 and 1;

[0047] 4) Perform density-based spatial clustering of the noise application to obtain the coordinates of all channel centers and channel radii;

[0048] 5) Count the number of aluminum atoms within a radius of c2 times the center of each channel to obtain the number of aluminum atoms in the intersection. Treat the aluminum atoms in the non-intersection of the bulk phase as the straight channel aluminum atoms; where c2 is a set constant greater than 1.

[0049] According to a second aspect of the present invention, an electronic device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the program to implement any of the methods described above.

[0050] According to a third aspect of the invention, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements any of the methods described herein.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] (1) The molecular sieve acid site placement simulation method designed in this invention can quickly and accurately simulate the placement of molecular sieve acid sites of different sizes and topologies under different temperatures and template agent conditions through giant canonical Monte Carlo simulation based on machine learning potential.

[0053] (2) In the process of giant canonical Monte Carlo simulation based on machine learning potential, the present invention adds high-precision density functional calculation results of molecular sieve-template agent to ensure relatively accurate extrapolation performance.

[0054] (3) The present invention uses the Gibbs free energy of the reaction to characterize the chemical potential change, and decides whether to accept the current structure according to the Metropolis criterion. This not only conforms to the thermodynamic laws to ensure the rationality of the simulation, but also allows for efficient exploration of the energy landscape by probabilistically accepting high-energy structures. Furthermore, it can quantitatively correlate energy and structural evolution, making it more flexible.

[0055] (4) The present invention adopts an efficient site identification method to accurately determine the placement of the template agent on the surface and bulk of molecular sieve nanoparticles. Through this method, the placement ratio of the template agent at different sites can be statistically analyzed, and the simulation results have stronger interpretability. Attached Figure Description

[0056] Figure 1 Flow chart of the method of the present invention.

[0057] Figure 2 This is a flowchart illustrating the specific operations in the embodiment.

[0058] Figure 3 The energy evolution curve of the giant canonical Monte Carlo (GCMC) in the embodiment is shown. Detailed Implementation

[0059] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0060] Example

[0061] like Figure 1 and Figure 2 As shown, this embodiment provides a method for simulating the placement of acidic sites in molecular sieves. The method includes the following steps:

[0062] S1. Generate all-silica molecular sieve nanoparticles with a set topology and size;

[0063] S2. Randomly insert a template molecule into the gap and select the nearest silicon atom to perform aluminum atom isomorphic substitution, as the initial guess structure of the molecular sieve particle-template adsorption state.

[0064] S3. Based on machine learning potential, obtain the relaxed structure of the adsorbed state and calculate the chemical potential change of the adsorption reaction;

[0065] S4. Based on the chemical potential change obtained in S3, decide whether to accept the current structure according to the Metropolis criterion;

[0066] S5. Update the current structure and go to S2 to generate a new initial guess structure based on the current structure. Repeat S2 to S5 and record the number of times the relaxed configuration is rejected consecutively. If the current number of consecutive rejections reaches the set threshold, end the simulation and go to S6.

[0067] S6. After determining the aluminum position, randomly remove a template agent molecule and randomly add hydrogen to any bridging oxygen connected to the framework aluminum closest to the original position of the template agent molecule. Based on the machine learning potential, obtain the relaxed structure and change the position of the hydrogen to connect with different bridging oxygens in the framework aluminum. Based on the machine learning potential, obtain the relaxed structure respectively. Determine the acid site position based on the lowest energy. Repeat S6 until all acid sites are in place.

[0068] S7. Based on the final molecular sieve-template structure, calculate and output the property parameters.

[0069] Next, using tetrapropylammonium cation as the synthetic template agent, the B acid site placement in ZSM-5 clusters with a particle size of 3 nm was simulated to provide a detailed explanation of the method of the present invention.

[0070] Initialization: Set global parameters such as cluster size, template agent and molecular sieve type, and reaction temperature.

[0071] S1. Generate all-silica molecular sieve nanoparticles with specific topological structures and sizes.

[0072] In this embodiment, the periodic structure of ZSM-5 was directly obtained from the International Zeolite Association database and expanded. The silicon atom closest to the center was selected, and all atoms within a distance of at least 1.5 nm from this silicon atom were counted. These atoms were then cut out of the periodic structure, and the information of their coordinating neighboring atoms was recorded. For atoms with neighboring atoms not in the cluster, their coordination characteristics were preserved as follows: if it was a silicon atom, a hydroxyl group was added as a ligand; if it was an oxygen atom, a hydrogen atom was added as a ligand.

[0073] S2. Randomly insert a template molecule into the gap and select the nearest silicon atom to perform aluminum atom isomorphic replacement, as the initial guess structure of the molecular sieve particle-template adsorption state.

[0074] Specifically, the Packmol input file is automatically generated, defining the task of randomly generating molecules within a specified region. There are two types of molecules: one is the current molecular sieve + template agent composite cluster, with a quantity of 1 (in the first step of the simulation, since no template agent molecule has been inserted yet, it is a pure molecular sieve nanoparticle generated in group 1); the other is the already modeled tetrapropylammonium cation, with a quantity of 1. The specified region is a spherical region with a radius of 2.0 nm (1.5 nm + 0.5 nm). The Packmol program is then called to generate the initial guessed structure.

[0075] S3. Call the machine learning potential function and LASP program trained based on LASP to automatically generate the LASP input file. Define the task as global optimization to obtain the relaxed configuration and energy, and calculate the Gibbs free energy of the reaction according to the following formula:

[0076]

[0077]

[0078] In the formula: To introduce the first n The chemical potential change of each template molecule; To determine the chemical potential after introducing the nth template molecule, The chemical potential before introducing the nth template molecule, E Represents the electronic energy term obtained based on the machine learning potential; C is a constant when the template agent and reaction temperature are selected, where: For newly inserted template Zero-point energy correction term; For the newly inserted template agent isobaric heat capacity integral correction term, For isobaric heat capacity, The reaction temperature, For newly inserted template Entropy correction term; The standard Gibbs free energy of formation for each species in a given state of matter and temperature in a corresponding chemical reaction; It is a hydroxide ion. Silicon, It is aluminum. Indicates the state of matter in an aqueous solution. It indicates the state of a solid substance.

[0079] Machine learning potential function (MLP) uses deep learning methods to fit quantum mechanical data, achieving an efficient approximation of the potential energy surface. This type of method has shown great potential in molecular dynamics. Based on the trained machine learning potential, this invention adds several high-precision density functional theory (DFT) calculation results from molecular sieve-template agents, thereby ensuring relatively accurate extrapolation performance.

[0080] In this embodiment, the architecture of the machine learning potential specifically includes:

[0081] 1) Machine learning can read the original structure of a species and the types and three-dimensional spatial coordinates of each atom;

[0082] 2) Encode the environment surrounding each atom using a centrosymmetric function to obtain a high-dimensional feature vector, where the i-th The high-dimensional eigenvectors corresponding to each atom can be represented as follows: ;

[0083] 3) Each high-dimensional feature vector is mapped to a partial energy value under the action of the same fully connected layer. The total energy of the system is equal to the sum of all the partial energies, expressed as:

[0084]

[0085] In the formula: It is a fully connected function, which is a mapping from a high-dimensional eigenvector space to the real number field.

[0086] This invention uses machine learning potential to perform giant canonical Monte Carlo simulations, which can simulate the placement of Brønsted acid sites in molecular sieves of different sizes and multi-level pore structures in open systems while ensuring the accuracy of thermodynamic quantity prediction.

[0087] S4. Based on the Gibbs free energy calculated in S3, the probability of accepting the relaxation configuration is:

[0088]

[0089] S5. Update the current structure and return to S2 to generate a new initial guess structure based on the current structure. Continue the loop of S2->S3->S4->S5->S2 and record the number of times the relaxed configuration is rejected consecutively in S4.

[0090] Set the threshold N_max for the number of consecutive rejections to 1000. If the number of rejections exceeds 1000, exit the loop.

[0091] S6. Perform the following cycle for all tetrapropylammonium cations in the current structure:

[0092] 1) Remove one tetrapropylammonium cation;

[0093] 2) Select the Al atom closest to the center N;

[0094] 3) Hydrogen is added to all the neighboring oxygen atoms bonded to the aluminum atom to generate multiple candidate sites for Brønsted acid;

[0095] 4) Use LASP to perform configurational relaxation based on machine learning potential to obtain the relaxed configuration and energy of all candidate sites;

[0096] 5) Select the configuration with the lowest relaxation energy as the current structure and update it, then return to 1) and continue the loop until all template agents are removed.

[0097] S7. Set the outermost layer to angstrom as the surface-bulk interface, count the number of silicon atoms and aluminum atoms on the surface and in the bulk, and obtain physical property parameters such as chemical composition, total silicon-aluminum ratio, bulk silicon-aluminum ratio, surface silicon-aluminum ratio, surface aluminum, and bulk aluminum. Obtain physical property parameters such as aluminum content in the straight channel and aluminum content in the intersection.

[0098] Specifically, it includes the following sub-steps:

[0099] S701, Identify aluminum in the surface and bulk phases by partition:

[0100] 1) Calculate and obtain the centroid of the molecular sieve nanoparticles;

[0101] 2) Take the maximum distance from the bulk atom to the center of mass as the radius of the cluster, and set the area within the cluster radius as the surface-bulk interface (5 angstroms in this embodiment), and count the number of silicon atoms and aluminum atoms on the surface and in the bulk.

[0102] S702. Identify aluminum in sections of straight channels and straight-channel-sinusoidal-channel intersections:

[0103] 1) Perform Voronoi partitioning on the cluster;

[0104] 2) Take the maximum distance from the bulk atom to the centroid as the radius of the cluster, and set the area within the cluster radius as the surface-bulk interface (5 angstroms in this embodiment), and obtain all Voronoi nodes in the bulk phase, as well as the distance from each node to the nearest atom;

[0105] 3) Obtain the maximum value of the distance from the Voronoi node to the nearest atom, with the maximum value as the upper bound and c1 times the maximum value as the lower bound, and retain all nodes whose distance to the nearest atom is between the upper and lower bounds; where c1 is a set constant from 0 to 1; in this embodiment, C1=0.9;

[0106] 4) Perform density-based spatial clustering of the noise application to obtain the coordinates of all channel centers and channel radii;

[0107] 5) Count the number of aluminum atoms within a radius of c2 times the center of each channel to obtain the number of aluminum atoms in the intersection. Take the aluminum atoms in the non-intersection of the bulk phase as the straight channel aluminum. Here, c2 is a set constant greater than 1. In this embodiment, c2=1.1.

[0108] Figure 3 The energy evolution curve of the giant canonical Monte Carlo (GCMC) in the embodiment is shown.

[0109] Table 1 below lists the physical property parameters obtained using the method of this embodiment, whose chemical composition is Si. 180 Al 48 O 540 H 100 TPA 48 TPA stands for tetrapropylammonium.

[0110] Table 1

[0111]

[0112] The electronic device of this invention includes a central processing unit (CPU), which can perform various appropriate actions and processes according to computer program instructions stored in read-only memory (ROM) or loaded from a storage unit into random access memory (RAM). The RAM may also store various programs and data required for device operation. The CPU, ROM, and RAM are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.

[0113] Multiple components in the device are connected to the I / O interface, including: input units such as keyboards and mice; output units such as various types of displays and speakers; storage units such as disks and optical discs; and communication units such as network interface cards (NICs), modems, and wireless transceivers. The communication unit allows the device to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0114] The processing unit executes the various methods and processes described above, such as methods S1 to S7. For example, in some embodiments, methods S1 to S7 may be implemented as computer software programs tangibly contained in a machine-readable medium, such as a storage unit. In some embodiments, part or all of the computer program may be loaded and / or installed on the device via ROM and / or a communication unit. When the computer program is loaded into RAM and executed by the CPU, one or more steps of methods S1 to S7 described above may be performed. Alternatively, in other embodiments, the CPU may be configured to execute methods S1 to S7 by any other suitable means (e.g., by means of firmware).

[0115] The functions described above in this document can be performed at least in part by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: field programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload programmable logic devices (CPLDs), and so on.

[0116] The program code used to implement the methods of the present invention can be written in any combination of one or more programming languages. This program code can be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code can be executed entirely on the machine, partially on the machine, as a standalone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0117] In the context of this invention, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0118] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and such modifications or substitutions are intended to be within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be subject to the scope of protection of the claims.

Claims

1. A method for simulating the acidic site placement of molecular sieves, characterized in that, include: S1. Generate all-silica molecular sieve nanoparticles with a set topology and size; S2. Randomly insert a template molecule into the gap and select the nearest silicon atom to perform aluminum atom isomorphic substitution, as the initial guess structure of the molecular sieve particle-template adsorption state. S3. Based on machine learning potential, obtain the relaxed structure of the adsorbed state and calculate the chemical potential change of the adsorption reaction; S4. Based on the chemical potential change obtained in S3, decide whether to accept the current structure according to the Metropolis criterion; S5. Update the current structure and go to S2 to generate a new initial guess structure based on the current structure. Repeat S2 to S5 and record the number of times the relaxed configuration is rejected consecutively. If the current number of consecutive rejections reaches the set threshold, end the simulation and go to S6. S6. After determining the aluminum position, randomly remove a template agent molecule and randomly add hydrogen to any bridging oxygen connected to the framework aluminum closest to the original position of the template agent molecule. Based on the machine learning potential, obtain the relaxed structure and change the position of the hydrogen to connect with different bridging oxygens in the framework aluminum. Based on the machine learning potential, obtain the relaxed structure respectively. Determine the acid site position based on the lowest energy. Repeat S6 until all acid sites are in place. S7. Based on the final molecular sieve-template structure, calculate and output the physical property parameters.

2. The method for simulating the acidic site placement of molecular sieves according to claim 1, characterized in that, The voids in S2 include the voids on the outer surface and the voids in the molecular sieve channels.

3. The method for simulating the acidic site placement of molecular sieves according to claim 1, characterized in that, The specific implementation of the machine learning potential in S3 includes: 1) Read the type and three-dimensional coordinates of each atom in the original structure; 2) Encode the environment surrounding each atom using a centrosymmetric function to obtain a high-dimensional feature vector; 3) Each high-dimensional feature vector is mapped to a partial energy value under the action of the same fully connected layer, realizing the mapping from the high-dimensional feature vector space to the real number field space. The final total energy is obtained by accumulating each partial energy value.

4. The method for simulating the acidic site placement of molecular sieves according to claim 1, characterized in that, The chemical potential change in S3 is characterized by the Gibbs free energy of the reaction, expressed as: The corresponding chemical reaction is: In the formula: To introduce the first n The chemical potential change of each template molecule; To introduce the first Chemical potential following a template molecule To introduce the first Chemical potential in front of a template molecule Represents the electronic energy term obtained based on the machine learning potential; C is a constant when the template agent and reaction temperature are selected, where: For newly inserted template Zero-point energy correction term; For the newly inserted template agent isobaric heat capacity integral correction term, For isobaric heat capacity, The reaction temperature, For newly inserted template Entropy correction term; The standard Gibbs free energy of formation for each species in a given state of matter and temperature in a corresponding chemical reaction; It is a hydroxide ion. Silicon, It is aluminum. Indicates the state of matter in an aqueous solution. It indicates the state of a solid substance.

5. The method for simulating the acidic site placement of a molecular sieve according to claim 1, characterized in that, In S4, the decision on whether to accept the current structure is made according to the Metropolis criterion, specifically as follows: 1) Calculate the probability of accepting the current structure, expressed as: In the formula: To introduce the first n The chemical potential change of each template molecule, Boltzmann's constant, The reaction temperature; 2) Generate a random number in the range [0,1], and the probability... Perform a size comparison; if the random number is smaller than... If the current structure is accepted, then the current structure is accepted; otherwise, the current structure is rejected.

6. The method for simulating the acidic site placement of a molecular sieve according to claim 1, characterized in that, In step S7, the outermost layer is set as the surface-bulk interface within angstroms. The number of silicon atoms and aluminum atoms in the surface and bulk phases are counted, and physical property parameters are calculated. The physical property parameters include chemical composition, total silicon-aluminum ratio, bulk silicon-aluminum ratio, surface silicon-aluminum ratio, surface aluminum, and bulk aluminum.

7. The method for simulating the acidic site placement of a molecular sieve according to claim 1, characterized in that, S6 specifically includes the following sub-steps: 1) After determining the location of aluminum, a template agent molecule is randomly removed, and hydrogen is added to all the bridging oxygens connected to the framework aluminum closest to the original position of the template agent molecule, generating multiple candidate acidic sites. 2) Based on machine learning potential, obtain the relaxed configuration and energy of all candidate points; 3) Select the configuration with the lowest energy after relaxation as the current structure and update it, then return to 1) and continue the cycle until all template agent molecules are removed.

8. The method for simulating the acidic site placement of a molecular sieve according to claim 1, characterized in that, S7, based on the final molecular sieve-template structure, calculates and outputs physical property parameters, specifically including the following sub-steps: S701, Identify aluminum in the surface and bulk phases by partition: 1) Calculate and obtain the centroid of the molecular sieve nanoparticles; 2) Take the maximum distance from the bulk atom to the center of mass as the radius of the cluster, and set the angstrom within the cluster radius as the surface-bulk interface, and count the number of silicon atoms and aluminum atoms on the surface and in the bulk. S702. Identify aluminum in sections of straight channels and straight-channel-sinusoidal-channel intersections: 1) Perform Voronoi partitioning on the cluster; 2) Take the maximum distance from the bulk atom to the centroid as the radius of the cluster, and set the angstrom within the cluster radius as the surface-bulk interface. Obtain all Voronoi nodes in the bulk phase, as well as the distance of each node to the nearest atom. 3) Obtain the maximum value of the distance from the Voronoi node to the nearest atom, with the maximum value as the upper bound and c1 times the maximum value as the lower bound, and retain all nodes whose distance to the nearest atom is between the upper and lower bounds; where c1 is a set constant between 0 and 1; 4) Perform density-based spatial clustering of the noise application to obtain the coordinates of all channel centers and channel radii; 5) Count the number of aluminum atoms within a radius of c2 times the center of each channel to obtain the number of aluminum atoms in the intersection. Treat the aluminum atoms in the non-intersection of the bulk phase as the straight channel aluminum atoms; where c2 is a set constant greater than 1.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the program, it implements the method as described in any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1 to 8.

Images