The COF Space: Materials Features, Gas Adsorption, and Separation Performances Assessed by Machine Learning

Area of Science:

• Materials Science
• Chemical Engineering
• Computational Chemistry

Background:

• Covalent organic frameworks (COFs) are emerging materials with significant potential for gas adsorption applications.
• Current research on COFs for gas separations is limited, hindering the exploration of their full capabilities.

Purpose of the Study:

• To comprehensively investigate the gas adsorption properties of a vast range of COFs (69,784 types) for critical industrial gas separations.
• To identify promising COF materials and understand the structure-property relationships governing their performance.

Main Methods:

• Utilized high-throughput molecular simulations to study the adsorption of CO2, CH4, H2, N2, and O2 in COFs across various pressures.
• Employed machine learning techniques to analyze adsorption data and predict COF performance.
• Evaluated COF performance for industrially relevant separations like CO2/CH4, CO2/H2, and O2/N2.

Main Results:

• Generated an extensive dataset of approximately 4.3 million data points covering structural, chemical, and energetic features of COFs.
• Identified key structural and chemical properties correlated with high adsorption capacity and selectivity for specific gases.
• Revealed the most promising COF adsorbents for critical gas separation processes.

Conclusions:

• This work provides the most extensive dataset to date for COF materials and their gas adsorption properties.
• The findings facilitate accelerated materials discovery for COFs in gas separation technologies.
• The study unlocks the potential of the broader COF material space for diverse applications.