Screening And Design Of Metal-organic Frameworks For Chiral Separation And Natural Gas Desulfurization

With the aggravation of energy crisis and environmental pollution,people pay more and more attention to the development and utilization of clean energy(such as natural gas).However,a large number of natural gas reservoirs remain undeveloped due to the presence of sulfides such as H₂S.Even a small proportion of H₂S will have a great adverse effect on the recovery and utilization of natural gas.Separation of acidic gas from natural gas is critical because the presence of acidic gas can lead to corrosion of separation equipment,production of harmful combustion products,and reduction of thermal efficiency.In pharmaceutical,agricultural and food industries,it is necessary and difficult to separate chiral molecules because only one of the two forms of chiral molecules is biologically active and the structure difference is very small.The design or selection of appropriate chiral stationary phase is the key to solve this problem.Metal-Organic Frameworks(MOFs)was adopted as an adsorbent,and high-throughput Computation(HTC)and Molecular fingerprint(MF)were used to explore the performance of MOFs applied to three systems(chiral separation of DMA molecules and removal of acidic gas in natural gas).HTC includes Grand Canonical Monte Carlo(GCMC)and Machine Learning(ML).ML is mainly represented by nonlinear algorithms such as Neighbor Component Analysis(NCA)and Gradient Boosting Regression Tree(GBRT).The MOFs were mainly from two databases,namely,Computation-Ready Experimental MOF(CoRE-MOF)and hypothetic MOF(h MOF).This study is about the chiral separation of DMA molecules and the removal of acidic gas in natural gas.The specific research is as follows:In order to solve the difficult problem of(R,S)-DMA chiral molecular separation,the design rules of new Functionalized homochiral MOFs(FHMOFs)with high enantioselectivity materials are explored.Firstly,45 FHMOFs were designed based on the random combination of S-KUMOF-1 and 10 functional groups.The Enantiomeric excess(ee)selectivity of FHMOFs at different temperatures was calculated by GCMC simulation.With increasing temperature,the ee for(R,S)-DMA are improved.The“glove effect”in the chiral pockets was proposed to explain the correlations between the steric effect of functional groups and performance of FHMOFs.Moreover,the neighborhood component analysis and RDKit/MACCS MFs show the highest predictive effect on enantioselectivities among the 4 ML classification algorithms with9 MFs that were tested.Based on the importance of MF,85 new FHMOFs were designed,and a newly designed FHMOF,NO₂-NHOH-FHMOF,with high similarity to the optimal MFs achieved improved chiral separation performance,with enantioselectivities of 85%.The design principles and new chiral pockets obtained by ML and MFs could facilitate the development of new materials for chiral separation.Aiming at the problem of removing acid gas in natural gas,starting from 606 CoRE-MOFs,34211 h MOFs were preliminarily designed using MF.TSN of 34211 h MOFs was increased to18.16,more than 4 times that of CoRE-MOFs,which represents a significant improvement in the performance of MOFs that designed by MF.In addition,the number of CoRE-MOFs with adsorption capacity greater than 1.0 mol/kg and TSN greater than 1.0 accounted for 5.44%.In the designed h MOFs,the number of MOF in this range accounts for 28.75%,which is more than five times that of CoRE-MOFs,indicating that the proportion of high-performance MOF that designed by MF is also greatly increased.Then,in order to reduce the interference of human factors,the program of molecular fingerprint-assisted automatic design of MOFs was developed.Due to a long time consuming,at present,only the first batch of h MOFs(23868)designed by this program was obtained.Comparisons of performance across different ranges show that high performance MOF ratio in the automatic design is more than three times higher than that in the initial design.Furthermore,the mean TSN of the three batches of MOF was0.380(606 CoRE-MOFs),0.757(34211 h MOFs),and 2.503(23868 h MOFs),respectively.Therefore,when the type of organic linker has a great influence on the performance of MOFs,using MF to design organic linker molecules can achieve the purpose of greatly improving the performance of new MOFs,which provides a design idea for calculators and experimenters.Finally,10 optimal h MOFs for removing acid gas(H₂S and CO₂)in natural gas were selected with the highest TSN up to 1600,100 times that of the original MOFs.At the same time,due to the shortcoming of time-consuming of molecular simulation in automatic design,an improvement strategy also was proposed to reduce the number and time of molecular simulation by establishing a prediction model based on ML.The prediction accuracy of GBRT algorithm can reach 0.96,theoretically satisfying the prediction.However,due to the shortcomings of the supervised algorithm,some excellent new MOFs(performance higher than the original data)will be ignored.Molecular simulation software RASPA and others were used to calculate the structure and adsorption properties of MOF and analyze the stability of the material,while Python and other software were used to analyze the data and develop automatic design programs.How to apply MF and ML to MOFs design is described in detail.The separation pattern of FHMOFs and the design rule of their chiral linkers were investigated by MF and ML,and three FHMOFs were designed to solve the difficult separation problem of DMA chiral molecules.Their comprehensive performance improved by 25%compared with the original FHMOFs.In addition,starting from CoRE-MOFs,more excellent h MOFs for acidic gas removal were gradually designed,and the comprehensive performance of the final designed h MOFs were more than six times that of CoRE-MOFs.ML and MF assisted GCMC can not only accelerate the speed of high-throughput calculation,but also dig deeper hidden information between material properties,composition and fingerprint,establish molecular design principles,and realize automatic design of high-performance MOF.This computational study provides microcosmic insights and design principles for experimental design,development and synthesis of novel MOF for efficient separation of chiral molecules and natural gas desulfurization,reducing experimental trial and error costs,and greatly improving the efficiency of experimental development of efficient MOF.