Integrating Molecular Simulations with Machine Learning Guides in the Design and Synthesis of [BMIM][BF₄]/MOF Composites for CO₂/N₂ Separation

Abstract

Considering the existence of a large number and variety of metal−organic frameworks (MOFs) and ionic liquids (ILs), assessing the gas separation potential of all possible IL/MOF composites by purely experimental methods is not practical. In this work, we combined molecular simulations and machine learning (ML) algorithms to computationally design an IL/MOF composite. Molecular simulations were first performed to screen approximately 1000 different composites of 1-n-butyl-3methylimidazolium tetrafluoroborate ([BMIM][BF 4 ]) with a large variety of MOFs for CO 2 and N 2 adsorption. The results of simulations were used to develop ML models that can accurately predict the adsorption and separation performances of [BMIM][BF 4 ]/MOF composites. The most important features that affect the CO 2 /N 2 selectivity of composites were extracted from ML and utilized to computationally generate an IL/MOF composite, [BMIM][BF 4 ]/UiO-66, which was not present in the original material data set. This composite was finally synthesized, characterized, and tested for CO 2 /N 2 separation. Experimentally measured CO 2 /N 2 selectivity of the [BMIM][BF 4 ]/UiO-66 composite matched well with the selectivity predicted by the ML model, and it was found to be comparable, if not higher than that of all previously synthesized [BMIM][BF 4 ]/MOF composites reported in the literature. Our proposed approach of combining molecular simulations with ML models will be highly useful to accurately predict the CO 2 /N 2 separation performances of any [BMIM][BF 4 ]/MOF composite within seconds compared to the extensive time and effort requirements of purely experimental methods.

MOF-GRU: A MOFid-Aided Deep Learning Model for Predicting the Gas Separation Performance of Metal–Organic Frameworks

“…For elements not covered by the DREIDING force field, we used parameters provided by the universal force field (UFF) . In previous research, the CH 4 adsorption and separation performance in MOFs simulated by the DREIDING + UFF force field demonstrated good agreement with the experimental data, even for MOFs with open metal sites. , This reliability has made it a preferred choice in various high-throughput screening studies of MOFs for CH 4 adsorption and separation applications. − Detailed parameter information can be found in Table S2. Each GCMC simulation cycle consisted of 1 × 10 6 equilibrium cycles and 1 × 10 6 production cycles for ensemble averaging.…”

Accelerated Discovery of Metal–Organic Frameworks for CO2 Capture by Artificial Intelligence

“…To find a novel IL/MOF composite for CO 2 /N 2 separation, ML models were developed using structural-, chemical-, and energy-based descriptors of 941 MOFs and their composites with [BMIM]­[BF 4 ] (1- n -butyl-3-methylimidazolium tetrafluoroborate) Figure (a) shows the distribution of ML-predicted CO 2 selectivities of [BMIM]­[BF 4 ]/MOF composites with respect to the two most important molecular features, porosity and pore volume, identified from the feature importance analysis of a ML model.…”

“…(a) 941 different types of [BMIM]­[BF 4 ]-incorporated IL/MOF composites and the relationships between their pore volume, porosity, and ML-predicted selectivities at 1 bar, 298 K. Readapted with permission from ref . (b) SHAP value analysis showing the relations between volumetric [MMIM]­[BF 4 ] loading (vol %) and ML-predicted TSN values for 15,410 different types of [MMIM]­[BF 4 ]-incorporated IL/COF composites at 1 bar, 298 K. Readapted with permission from ref .…”

Machine learning-assisted computational exploration of the optimal loading of IL in IL/COF composites for carbon dioxide capture

“…35 To the best of our knowledge, among all the existing studies, the content of IL in porous materials is generally described by gravimetric loading ratio. 36–40 However, such wt% parameter cannot describe the degree of modification on the pore environment ( i.e. , pore volume and/or pore size).…”