Data-Driven Evaluation as a Preliminary Tool to Judiciously Choose Covalent Organic Frameworks to be used as Fillers in Mixed Matrix Membranes for Hydrogen Separation

Hydrogen is considered the fuel of the future for the decarbonizing industry and the transport sector. A membrane containing porous materials can be used to extract hydrogen from various gas mixtures. The diversity of porous materials makes it difficult to select a specific material for hydrogen separation. Covalent organic frameworks (COFs) are becoming increasingly popular for gas separation processes, especially for high-temperature applications. The properties of a COF depend on the configuration of various building blocks. A data-driven evaluation was performed to select the most promising COF features and topologies from the CoRE COF database of 1242 reported COFs based on quantification using a scoring approach. The results of the property-based evaluation model show that 3D COFs have a higher potential on average than 2D COFs. Four different normalized evaluations describe the design with the highest potential. The interdisciplinary quantification approach of the preferred properties enables the selection of a specific topology from 2D or 3D covalent organic frameworks and serves as an indication of further advantageous features for a selective hydrogen separation material.