Machine Learning Prediction of Henry Coefficients of Polar and Nonpolar Gases in Covalent Organic Frameworks: Effects of Interlayer Shifts and Functionalization

Covalent organic frameworks (COFs) are promising materials for gas separation and carbon capture. Computational techniques based on Monte Carlo simulation can be used to predict the gas adsorption properties of COFs with high accuracy, however they are too inefficient to be deployed in a high-throughput manner for screening large COF databases. In this paper, we systematically train and evaluate a range of machine learning models for predicting the Henry coefficients for CO2 and CH4 gas adsorption in COF materials. To account for COF structural variability, we train our models on datasets that include both chemically functionalized frameworks and interlayer displaced stacking configurations. By comparing predictive performance across descriptor–model architecture combinations, we demonstrate how different models capture the key physical factors governing gas adsorption, including electrostatics, local atomic environments, and van der Waals interactions. Our results therefore provide a framework for building machine learning models for scalable, high-throughput screening of COF materials with targeted gas adsorption properties.