Rapid and Accurate Screening of the COF Space for Natural Gas Purification: COFInformatics

Abstract

In this work, we introduced COFInformatics, a computational approach merging molecular simulations and machine learning (ML) algorithms, to evaluate all synthesized and hypothetical covalent organic frameworks (COFs) for the CO 2 /CH 4 mixture separation under four different adsorptionbased processes: pressure swing adsorption (PSA), vacuum swing adsorption (VSA), temperature swing adsorption (TSA), and pressure−temperature swing adsorption (PTSA). We first extracted structural, chemical, energy-based, and graph-based molecular fingerprint features of every single COF structure in the very large COF space, consisting of nearly 70,000 materials, and then performed grand canonical Monte Carlo simulations to calculate the CO 2 /CH 4 mixture adsorption properties of 7540 COFs. These features and simulation results were used to develop ML models that accurately and rapidly predict CO 2 /CH 4 mixture adsorption and separation properties of all 68,614 COFs. The most efficient separation process and the best adsorbent candidates among the entire COF spectrum were identified and analyzed in detail to reveal the most important molecular features that lead to high-performance adsorbents. Our results showed that (i) many hypoCOFs outperform synthesized COFs by achieving higher CO 2 /CH 4 selectivities; (ii) the top COF adsorbents consist of narrow pores and linkers comprising aromatic, triazine, and halogen groups; and (iii) PTSA is the most efficient process to use COF adsorbents for natural gas purification. We believe that COFInformatics promises to expedite the evaluation of COF adsorbents for CO 2 /CH 4 separation, thereby circumventing the extensive, time-and resource-intensive molecular simulations.

Simulation and Machine Learning-Integrated Investigation of Double-Linker MOFs for the Separation of CF4 from CH4 and N2

“…It should be noted that in the literature, many studies considered similar lists of features to those that are used in this study. ,− However, the feature list could involve different features such as atom density, energy histograms, atomic property weighted radial distribution functions (AP-RDFs), revised autocorrelation functions (RACs), etc., which are being increasingly used. Both the features involved and not involved in this study can be used not only for MOFs but also for other porous materials such as covalent–organic frameworks (COFs) and zeolites. − It should be noted that, in contrast to this work, some of the studies in the literature involve ML models with more than two features to predict gas adsorption/separation performances. , In those ML models, many different features such as topology, void fraction, functional group density, most positive charge, most negative charge, metal angle, and surface atom density were utilized. Depending on the adsorbates, materials, adsorption conditions, and target values, different features could be the most important feature.…”