Machine learning using host/guest energy histograms to predict adsorption in metal–organic frameworks: Application to short alkanes and Xe/Kr mixtures

A machine learning (ML) methodology that uses a histogram of interaction energies has been applied to predict gas adsorption in metal–organic frameworks (MOFs) using results from atomistic grand canonical Monte Carlo (GCMC) simulations as training and test data. In this work, the method is first extended binary mixtures spherical species, particular, Xe Kr. addition, it shown single-component ethane propane can be predicted good agreement with GCMC simulation felt by methyl probe conjunction random forest ML method. The for improved including small number MOF textural properties descriptors. We also discuss most significant features, which provides physical insight into beneficial energy sites given application.