A screening method and system for molecular sieves for removing trace phosphorus impurities from hydrogen

Abstract

The present invention discloses a method and system for screening molecular sieves for removing trace phosphorus impurities in hydrogen. The method steps include: developing a force field based on the DFT method to obtain a simulated force field, which is used to describe the adsorption capacity of the molecular sieve for trace phosphorus impurities; based on the simulated force field, obtaining an optimal molecular sieve structure, and the optimal molecular sieve is used to remove trace phosphorus impurities in hydrogen. The present invention selects a cluster model as the basic model for force field adjustment based on multiple cluster models. Compared with the previous force field adjustment methods, it can save more computational costs. It has great reference significance for the simulation and screening of gases without experimental data and with high toxicity. At the same time, the adsorption behavior of trace PH3 in the recycled hydrogen in the all-silica molecular sieve can be accurately predicted by simulation means, thus avoiding as much as possible the large amount of manpower and material resources required for experimental operations.

Description

Technical Field

[0001] The present invention relates to the fields of computational chemistry and polysilicon production, and particularly relates to a method and system for screening molecular sieves for removing trace phosphorus impurities from hydrogen gas. Background Art

[0002] As a multifunctional material, polysilicon is widely used in the fields of photoelectric conversion, information processing, and artificial intelligence due to its good electrical conductivity and unique optical properties. The content of donor and acceptor impurities represented by boron and phosphorus impurities is an important index affecting the grade evaluation of polysilicon. More than 85% of the global polysilicon products are produced by the improved Siemens method, and excessive hydrogen gas is recycled throughout the production process. Even if there is only a trace amount of PH3 impurity in the hydrogen gas, it will continuously accumulate with the production cycle, thereby affecting the quality of polysilicon products. Therefore, designing and developing new adsorbents to achieve deep removal of trace PH3 in recycled hydrogen gas is crucial for the entire polysilicon production field.

[0003] In recent years, molecular sieves have been widely used in the field of gas adsorption and separation due to their high specific surface area, large adsorption capacity, and excellent adsorption selectivity and regeneration performance. Existing research has proven that different molecular sieve topologies have a greater impact on the removal efficiency of PH3 in recycled hydrogen gas. If experimental screening is used to select the molecular sieve topology suitable for PH3 removal, hundreds of all-silica molecular sieves need to be prepared and multiple adsorption experiments need to be carried out. Such a screening method is relatively blind, cumbersome, costly, and has a long experimental cycle. Summary of the Invention

[0004] To solve the above technical problems in the background, the present invention aims to provide a simple force field development method and apply this method to the screening of molecular sieves for removing trace phosphorus impurities from hydrogen gas. The above method is aimed at toxic and harmful substances that have no large amount of experimental data support and have a content of only ppm level in the adsorption main phase. Due to the lack of experimental data, the force field parameters that can accurately describe their adsorption behavior cannot be obtained, and thus the screening of molecular sieve topologies cannot be carried out through GCMC simulation.

[0005] To achieve the above object, the present invention provides a method for screening molecular sieves for removing trace phosphorus impurities from hydrogen gas, and the steps include:

[0006] Based on the DFT method, force field development is carried out to obtain a simulated force field, and the simulated force field is used to describe the adsorption capacity of the molecular sieve for trace phosphorus impurities;

[0007] Based on the simulated force field, an optimal molecular sieve structure is obtained, and the optimal molecular sieve structure is used to remove trace phosphorus impurities from hydrogen gas.

[0008] Preferably, the method for carrying out the force field development includes:

[0009] Statistically analyze the cyclic structural units of all-silica molecular sieves in the International Zeolite Database, and intercept the molecular sieve cluster model according to the occurrence frequency;

[0010] Based on the molecular sieve cluster model, obtain a number of potential adsorption configurations;

[0011] Based on the potential adsorption configurations, construct a preliminary simulation force field;

[0012] By adjusting the existing molecular sieve force field parameters and calculating the mutual binding energy, adjust the preliminary simulation force field based on the binding energy calculation results of density functional theory to obtain the simulation force field.

[0013] Preferably, the method for intercepting the molecular sieve cluster model includes: statistically analyzing the cyclic structural units of all-silica molecular sieves in the International Zeolite Database, statistically analyzing the cyclic structural units, and selecting the one with the highest occurrence frequency as the molecular sieve cluster model.

[0014] Preferably, the method for adjusting the preliminary simulation force field includes: based on the Trappe-Zeo force field of existing molecular sieves, adjust the Lennard-Jones force field parameters of the Trappe-Zeo force field by continuously amplifying or reducing the potential energy and well depth of the L-J parameters, so that the calculated mutual binding energy gradually approaches the DFT calculation results of Gaussian until it is consistent with the mutual binding energy obtained by the DFT calculation method.

[0015] Preferably, the method for performing the screening includes: performing GCMC simulation on the removal of trace PH3 in a hydrogen atmosphere using the parameters of the simulation force field to obtain the optimal structural parameter range of the molecular sieve and the optimal molecular sieve structure.

[0016] The present invention also provides a screening system for molecular sieves for removing trace phosphorus impurities in hydrogen, which is used to implement the above method, including: a development module and a screening module;

[0017] The development module is used to perform force field development based on the DFT method to obtain a simulation force field;

[0018] The screening module is used to obtain the optimal molecular sieve structure based on the simulation force field.

[0019] Preferably, the working process of the development module includes:

[0020] Statistically analyze the cyclic structural units of all-silica molecular sieves in the International Zeolite Database, and intercept the molecular sieve cluster model according to the occurrence frequency;

[0021] Based on the molecular sieve cluster model, obtain a number of potential adsorption configurations;

[0022] Based on the potential adsorption configuration, a preliminary simulation force field is constructed;

[0023] By adjusting the existing molecular sieve force field parameters and calculating the binding energy, the preliminary simulation force field is adjusted based on the binding energy calculation results of density functional theory to obtain the simulation force field.

[0024] Preferably, the working process of the screening module includes: performing GCMC simulation on the removal of trace PH3 in a hydrogen atmosphere using the parameters of the simulation force field to obtain the optimal structural parameter range of the molecular sieve and the optimal molecular sieve structure.

[0025] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0026] The present invention selects a cluster model as the basic model for force field adjustment based on multiple cluster models. Compared with the previous force field adjustment methods, it can save more calculation costs. It has great reference significance for the simulation and screening of gases without experimental data and with high toxicity. At the same time, the adsorption behavior of trace PH3 in recycled hydrogen in all-silica molecular sieves can be accurately predicted by simulation means, avoiding as much as possible the large consumption of manpower and material resources required for experimental operations. In addition, the present invention can also summarize the adsorption behavior of PH3 by the topological structure of all-silica molecular sieves in the IZA molecular sieve database to obtain the characteristic structural parameter range suitable for PH3 adsorption. Larger pore intersections in the molecular sieve can better capture PH3 molecules. The above findings can provide guidance for the further modification of molecular sieves. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the technical solutions of the present invention, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0028] Figure 1 It is a schematic flowchart of the method of the embodiment of the present invention;

[0029] Figure 2 It is a verification diagram of the CH4 adsorption isotherm of the MFI molecular sieve framework at 303K before and after the force field adjustment of CH4-all-silica molecular sieve in the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0030] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0031] To make the above objects, features, and advantages of the present invention more obvious and understandable, the present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0032] Before the description, first introduce the currently more mainstream molecular sieve screening methods.

[0033] In recent years, through molecular simulation or machine learning means, using the grand canonical Monte Carlo (GCMC) method to simulate the adsorption of target adsorbate molecules by various adsorbents to screen ideal adsorbents has become a powerful approach in adsorption research.

[0034] However, for GCMC simulation, the correctness of the force field determines the reliability of the simulation results. Generally speaking, the reliability of the force field will be verified through experimental data. However, the traditional first-principles force field development method is time-consuming, and gases such as PH3 are highly toxic, with only ppm-level content in a hydrogen atmosphere, making it difficult to conduct experiments for measuring adsorption isotherms and adsorption heats, and it is difficult to verify the correctness of existing force field parameters.

[0035] Therefore, aiming at the above-mentioned defects of the existing technology, the present invention simplifies the development method of the existing force field, develops a new force field through a method verified by the mutual binding energy with DFT calculation, and uses the developed force field for the screening of the molecular sieve topology in the removal of trace phosphorus impurities in hydrogen. The steps are as Figure 1 shown.

[0036] Example 1

[0037] Next, in conjunction with this embodiment, it will be detailed how the present invention solves technical problems in real life.

[0038] Before the force field development, it is necessary to optimize the structure of the adsorbate molecules to ensure the correctness of calculating the mutual binding energy. All DFT calculations during the optimization process are implemented using Gaussian, using the B3LYP, 6-311+G(d, p) basis set, and the influence of dispersion forces is considered during the calculation process.

[0039] S1. Develop a force field based on the DFT method to obtain a simulated force field. The specific steps include:

[0040] Statistically analyze the cyclic structural units of all-silica molecular sieves in the International Zeolite Association (IZA) database. Intercept the molecular sieve cluster model according to the occurrence frequency; specifically, statistically analyze the basic ring-shaped composition units of the all-silica molecular sieve topology, and select the ring-shaped model with the highest occurrence frequency as the molecular sieve cluster model. In this embodiment, the final molecular sieve cluster model is selected as the six-membered ring and the eight-membered ring; and the initial force field of the all-silica molecular sieve is selected as the Trappe-Zeo force field.

[0041] After that, by continuously adjusting the distance between the adsorbate molecule and the molecular sieve cluster model, several potential adsorption configurations are obtained. Among them, the distance between the six-membered ring cluster model and the PH3 molecule is set as follows: The distance between the eight-membered ring cluster model and the PH3 molecule is set as follows: Form a preliminary simulation force field. During the development of the approximate adsorption model, the cluster model and the PH3 molecule cannot intersect or overlap. These potential adsorption models will be used as data for subsequent steps to prepare for better adjustment of the force field parameters closer to the real adsorption.

[0042] Finally, by adjusting the existing molecular sieve force field parameters and calculating the binding energy, the preliminary simulation force field is adjusted based on the binding energy calculation results of density functional theory to obtain the simulation force field. Specifically, the PH3 molecule adopts the UFF force field. Based on the existing Trappe-Zeo force field of the molecular sieve, the Lennard-Jones force field parameters of the Trappe-Zeo force field are adjusted by using the Forcite module in Material Studio to continuously magnify or reduce the potential energy and well depth of the L-J parameters, so that the calculated binding energy gradually approaches the DFT calculation result of Gaussian until it is consistent with the binding energy obtained by the DFT calculation method. To ensure the correctness of the binding energy calculated by molecular dynamics, the bond lengths and bond angles in the PH3 molecule must be set to accurate values during the calculation and adjustment of the force field parameters. Specifically, the bond length of the P-H bond is The bond angle is 93.52°.

[0043] In this embodiment, the entire process of force field adjustment is carried out in the Forcite module of Material Studio. The calculation method of the binding energy between the Forcite modules is different from the DFT calculation, and the force field can be verified while establishing the force field.

[0044] Table 1 lists the force field adjustment process of PH3 on the zeolite cluster model, and Table 2 represents the calculated values of the binding energy by two calculation methods under the optimal force field parameters. It can be seen that when the potential energy ε and well depth σ of the L-J parameters of the all-silica zeolite are expanded to 1.2 times and 1.1 times of the original respectively, the relative error of the binding energy obtained by the two calculation methods for each cluster model is the smallest. The force field parameters at this time are considered to be able to accurately describe the adsorption capacity of PH3 on the all-silica zeolite.

[0045] Table 1

[0046]

[0047] Table 2

[0048]

[0049] S2. Based on the simulated force field, the zeolite topologies are screened to obtain the optimal zeolite.

[0050] Based on the adjusted parameters of the simulated force field, the adsorption of PH3 by 103 all-silica zeolites in the IZA zeolite database is simulated using the RASPA software. The adsorption capacity of the zeolite framework for PH3 is statistically analyzed to obtain the range of the optimal zeolite structure parameters and the optimal zeolite topology most suitable for PH3 removal.

[0051] First, the zeolite structure parameters are calculated, including zeolite structure parameters such as PLD (restricted pore diameter), LCD (maximum cavity diameter), ASA (specific surface area), AV (pore volume), φ (porosity), and R (PLD / LCD). 103 zeolite topologies with PLD greater than the molecular dynamic diameter of PH3 and continuous structures are pre-screened; at the same time, the spaces in the zeolite structure that PH3 cannot enter are blocked. When calculating the structure parameters, the preferred probe molecule is the helium molecule, with a diameter of The blocking of the inaccessible area of the zeolite uses PH3 as the probe molecule to prevent the GCMC simulation method from directly inserting PH3 molecules into the inaccessible zeolite space.

[0052] It should be noted that the calculation of the zeolite structure parameters in this embodiment is only for the convenience of preliminarily screening the zeolite structure according to the size of the adsorbate molecules and preparing for the next step of optimizing the range of the optimal zeolite structure parameters. The calculation software includes but is not limited to Zeo++, MERCURY, Material Studio, etc.

[0053] After that, under conditions close to industrial conditions, GCMC adsorption simulations of trace PH3 molecules in a hydrogen atmosphere were carried out for different zeolite structures. The zeolite structure parameters most suitable for PH3 capture were obtained, and 5 zeolite topologies with excellent adsorption capacity and adsorption selectivity were selected. The specific simulation conditions were -15 °C and 900 kPa, and the concentration of PH3 in the circulating hydrogen was 100 ppm.

[0054] It should be noted that in this embodiment, the GCMC simulation is only for facilitating the screening of the optimal zeolite topology structure and the optimal parameter range according to the developed force field. The simulation software includes but is not limited to RASPA and Material Studio.

[0055] Finally, with the adsorption amount of PH3 by the zeolite during single-component adsorption and competitive adsorption as the abscissa and ordinate respectively, 5 zeolite topologies of STI, IMF, OKO, TUN, and MEL with relatively large adsorption selectivity and adsorption capacity were selected.

[0056] The method for determining the optimal zeolite structure parameter range includes: taking the corresponding structure parameters of 103 zeolites as the abscissa and the magnitude of the adsorption capacity of the zeolite for PH3 during competitive adsorption as the ordinate, and obtaining the optimal value within the range of the first 5% and the last 5% of the optimal value. The selected optimal zeolite structure parameter range includes:

[0057] (Pore channel restricted diameter) (Maximum pore cavity diameter) 450m 2 / g < ASA (Specific surface area) < 550m 2 / g, 0.17cm 3 / g < AV (Pore volume) < 0.23cm 3 / g, 0.25 < φ (Porosity) < 0.35, 0.73 < R (PLD / LCD) < 0.83.

[0058] Example 2

[0059] To verify the accuracy of the force field of the present invention, in this embodiment, CH4 molecules were specifically selected as the experimental object for verification; the reason for selecting CH4 molecules is that their non-toxic characteristics enable them to have very complete DFT data, and the verification results are as Figure 2 shown.

[0060] First, the basic building units of six-membered rings and eight-membered rings in classical zeolites were intercepted, and CH4 molecules were placed near the six-membered rings and eight-membered rings. The distance between the adsorbate molecules and the cluster model was adjusted to obtain two approximate adsorption models for each of the six-membered ring and eight-membered ring. The distances between the six-membered ring cluster model and the CH4 molecules were respectively set to: Set the distance between the eight-membered ring cluster model and the CH4 molecule to:

[0061] Afterwards, based on the Trappe-Zeo force field of molecular sieves, the Forcite module in Material Studio was used to continuously adjust the Lennard-Jones force field parameters of the Trappe-Zeo force field so that the calculated binding energy was close to the DFT calculation results of Gaussian.

[0062] Table 3 lists the force field adjustment process of CH4 on the molecular sieve cluster model. Table 4 represents the calculated values ​​of binding energy by the two calculation methods under the optimal force field parameters. It can be seen that when the LJ parameter potential energy and well depth of the all-silicon molecular sieve are expanded to 1.2 times and 1.05 times of the original, respectively, the relative error of the binding energy obtained by the two calculation methods for each cluster model is the smallest. The force field parameters at this time are considered to accurately describe the adsorption capacity of CH4 on all-silicon molecular sieves. Furthermore, RASPA can reproduce the adsorption isotherm of CH4 in MFI molecular sieves at 303K, which fully proves the feasibility of this LJ force field adjustment method and screening method in screening molecular sieve structures.

[0063] Table 3

[0064]

[0065] Table 4

[0066]

[0067] Embodiment 3

[0068] This embodiment is the same as embodiment 1, except that the adsorbate molecule is replaced by HCl from PH3.

[0069] (1) The basic components of the molecular sieve, the six-membered ring and the eight-membered ring, are intercepted, the HCl molecule is placed near the six-membered ring and the eight-membered ring, and the distance between the adsorbate molecule and the cluster model is adjusted to obtain two approximate adsorption models of the six-membered ring and the eight-membered ring. The distance between the six-membered ring cluster model and HCl is set to: Set the distance between the eight-membered ring cluster model and HCl to:

[0070] (2) Based on the Trappe-Zeo force field of molecular sieves, the Lennard-Jones force field parameters of the Trappe-Zeo force field were continuously adjusted using the Forcite module in Material Studio so that the calculated binding energy was close to the DFT calculation result of Gaussian.

[0071] Table 5 lists the force field adjustment process of HCl molecules on the zeolite cluster model. Table 6 represents the calculated values of the binding energy by two calculation methods under the optimal force field parameters. It can be seen that when the L-J parameter potential energy and well depth of the all-silica zeolite are expanded to 1.4 times and 1.05 times of the original respectively, the relative error of the binding energy obtained by the two calculation methods for each cluster model is the smallest.

[0072] Table 5

[0073]

[0074] Table 6

[0075]

[0076] From the above Examples 1 to 3, it can be concluded that the zeolite topology screening method provided by the present invention can be applied to the simulation of gas adsorption and separation of various highly toxic gases and gases without existing force field parameters. It is difficult to obtain experimental parameters for such gases, and it is impossible to compare the relevant force field parameters obtained by traditional methods with the experimental results. It can be seen from Example 2 that the simulation data obtained by this zeolite topology screening method can fit well with the experimental values, proving that the force field can also be verified when experimental data is relatively lacking, which has certain practical significance. It can be seen from Examples 1 to 3 that this force field adjustment method is relatively simple and has a lower cost compared with the traditional force field development method.

[0077] Example 4

[0078] This example also provides a screening system for zeolites for removing trace phosphorus impurities in hydrogen, including: a development module and a screening module; the development module is used to develop a force field based on the DFT method to obtain a simulated force field; the screening module is used to obtain the optimal zeolite structure based on the simulated force field.

[0079] The working process of the development module includes: counting the cyclic structural units of all-silica zeolites in the international zeolite database, intercepting the zeolite cluster model according to the appearance frequency; obtaining a number of potential adsorption configurations based on the zeolite cluster model; constructing a preliminary simulated force field based on the potential adsorption configurations; adjusting the preliminary simulated force field by adjusting the existing zeolite force field parameters and calculating the mutual binding energy, and obtaining the simulated force field based on the binding energy calculation results of the density functional theory.

[0080] The working process of the screening module includes: performing GCMC simulation on the removal of trace PH3 in a hydrogen atmosphere using the parameters of the simulated force field to obtain the optimal structural parameter range of the zeolite and the optimal zeolite structure.

[0081] The embodiments described above are only descriptions of the preferred embodiments of the present invention, and do not limit the scope of the present invention. Without departing from the design spirit of the present invention, various deformations and improvements made by those of ordinary skill in the art to the technical solutions of the present invention shall fall within the protection scope determined by the claims of the present invention.

Claims

1. A screening method for molecular sieves for removing trace phosphorus impurities in hydrogen, characterized in that the steps Including: Develop a force field based on the DFT method to obtain a simulated force field, which is used to describe the adsorption capacity of molecular sieves for trace phosphorus impurities; Based on the simulated force field, obtain the optimal molecular sieve structure, which is used to remove trace phosphorus impurities in hydrogen; The method for developing the force field includes: counting the cyclic structural units of all-silica molecular sieves in the international molecular sieve database, and intercepting the molecular sieve cluster model according to the occurrence frequency; based on the molecular sieve cluster model, obtain several potential adsorption configurations; based on the potential adsorption configurations, construct a preliminary simulated force field; by adjusting the existing molecular sieve force field parameters and calculating the mutual binding energy, adjust the preliminary simulated force field based on the binding energy calculation results of density functional theory to obtain the simulated force field; The method for intercepting the molecular sieve cluster model includes: by counting the cyclic structural units of all-silica molecular sieves in the international molecular sieve database, statistically analyzing the cyclic structural units, and selecting the one with the highest occurrence frequency as the molecular sieve cluster model; The method for adjusting the preliminary simulated force field includes: based on the Trappe-Zeo force field of the existing molecular sieve, adjust the Lennard-Jones force field parameters of the Trappe-Zeo force field by continuously enlarging or shrinking the potential energy and well depth of the L-J parameters, so that the calculated mutual binding energy gradually approaches the DFT calculation result of Gaussian until it is consistent with the mutual binding energy obtained by the DFT calculation method.

2. The screening method of the molecular sieve for removing trace phosphorus impurities in hydrogen according to claim 1, wherein The method for performing the screening includes: performing GCMC simulation on the removal of trace PH3 in a hydrogen atmosphere using the parameters of the simulated force field to obtain the optimal structural parameter range of the molecular sieve and the optimal molecular sieve structure.

3. A screening system for molecular sieves to remove trace phosphorus impurities from hydrogen, the system being used to implement the method according to any one of claims 1-2, characterized in that, Including: A development module and a screening module; The development module is used to develop a force field based on the DFT method to obtain a simulated force field, which is used to describe the adsorption capacity of molecular sieves for trace phosphorus impurities; The working process of the development module includes: counting the cyclic structural units of all-silica molecular sieves in the international molecular sieve database, and intercepting the molecular sieve cluster model according to the occurrence frequency. The process includes counting the cyclic structural units of all-silica molecular sieves in the international molecular sieve database, statistically analyzing the cyclic structural units, and selecting the one with the highest occurrence frequency as the molecular sieve cluster model; based on the molecular sieve cluster model, obtain several potential adsorption configurations; based on the potential adsorption configurations, construct a preliminary simulated force field; by adjusting the existing molecular sieve force field parameters and calculating the mutual binding energy, adjust the preliminary simulated force field based on the binding energy calculation results of density functional theory to obtain the simulated force field. The process includes based on the Trappe-Zeo force field of the existing molecular sieve, adjust the Lennard-Jones force field parameters of the Trappe-Zeo force field by continuously enlarging or shrinking the potential energy and well depth of the L-J parameters, so that the calculated mutual binding energy gradually approaches the DFT calculation result of Gaussian until it is consistent with the mutual binding energy obtained by the DFT calculation method; The screening module is used to obtain the optimal molecular sieve structure based on the simulated force field, and the optimal molecular sieve structure is used to remove trace phosphorus impurities in hydrogen.

4. The screening system for molecular sieves for removing trace phosphorus impurities from hydrogen gas according to claim 3, characterized in that, The working process of the screening module includes: performing GCMC simulation on the removal of trace PH3 in a hydrogen atmosphere by using the parameters of the simulated force field to obtain the range of optimal structural parameters of the molecular sieve and the optimal molecular sieve structure.

Images