High-throughput Molecular Simulation Study Of Porous Material Membranes For Separation And Purification Of H₂ And CO₂

After the second industrial revolution,fossil fuels became the main energy source in industrial development.While accelerating the development of science and technology and changing the way of life of human beings,fossil fuels also bring two major problems that need to be solved urgently:the greenhouse effect and the crisis of fossil energy.Gas separation technology is an effective solution.Separating H₂ can replace fossil fuels to solve the energy crisis,and separating CO₂ for storage can effectively alleviate environmental problems.Metal-organic framework（MOF）materials show broad application prospects in the field of gas separation due to their large porosity,tunable pore size,and excellent reproducibility.The combination of MOF materials and inorganic particles,a mixed matrix membrane（MMM）,can be used to push the upper limit of gas separation.Considering that the number of MOFs is increasing year by year,and the number of MMMs is also increasing,it is time-consuming and laborious to test the performance of MOFs and MMMs one by one through experiments.Therefore,based on the MOF database and the derived MMM database,this study adopts high-throughput computation（HTC）and machine learning（ML）methods for the gas separation of H₂ gas mixture system and CO₂ gas mixture system.A systematic theoretical study has been carried out,and the specific research contents are as follows:1.The separation performance of 6013 computationally-ready,experimental metal-organic framework membrane（Co RE-MOFM）were evaluated using two commonly used molecular simulation methods:grand canonical Monte Carlo（GCMC）and molecular dynamic（MD）.We screened out a series of optimal MOF candidates for 6 H₂ mixtures（H₂/X,X=CH₄,N₂,H₂S,O₂,CO₂,He）,based on the results of the structure-performance relationships by univariate analysis found that the performance of the top-performing candidates far exceeded Robeson’s upper bound.The MOFM structure-property relationship was revealed by univariate analysis.To trade off permeability P_H2 and permselectivity S_perm,i/j,a new trade-off value—trade-off of multiple selectivity and permeability（TMSP）—is defined as the comprehensive performance metric of Co RE-MOFM.It was found that TMSP can effectively evaluate the separation performance of Co RE-MOFM for various H₂ mixtures.In order to speed up the computational screening process,8 ML methods based on supervised learning algorithms were used to predict P_H2 and TMSP,and the results showed that Gaussian process regression（GPR）and random forest（RF）methods have high predictive ability.After calculating each feature importance and using a feature selection method,the relationship between the number of features and the accuracy of the ML model was explored.Finally,a decision tree model based on the principle of the minimum Gini coefficient is built,which accurately classifies Co RE-MOFM through the material descriptor path.2.The introduction of MOF into the polymer matrix can effectively improve the gas separation performance of the mixed matrix membrane（MMM）.11 kinds of polymers were selected,including 2 kinds of polymers of intrinsic microporosity（PIM-1,PIM-7）,amorphous polyethylene oxide,semi-crystalline polyethylene oxide Ethylene,Polyphenylene Oxide,Cellulose acetate,3 kinds of polyimide（Matrimid,PI-3,PI-5）,poly[1-（trimethylsilyl）-1-propyne],poly(4-methyl-1-pentene,combined with 6013 MOFs with different volume fractions（0.1,0.2,0.3）into nearly 200,000 MMMs.The gas separation performance of the MMM was calculated using the Maxwell equation,and how the relationship between MOF and polymer properties affected the MMM performance was explored.This study combines HTC and ML to deeply explore the relationship between MOF characteristics,gas properties,and gas separation performance.The quantitative structure-performance relationship revealed at the microscale can provide a theoretical basis for screening high-performance Co RE-MOFM,and the HTCS method and ML algorithm can efficiently accelerate the design and development of high-performance H₂-separated Co RE-MOFM.This study aims to provide experimenters with directions for material synthesis,reducing trial and error costs.These strategies for screening and discovering materials provide guidance for accelerating the development of H₂ and CO₂ separation membranes,and ultimately achieve the ultimate goal of energy efficiency and environmental cleanliness.