In Silico Study of Metal-Organic Frameworks for CO2/CO Separation: Molecular Simulations and Machine Learning

Metal-organic frameworks (MOFs), an emerging classof nanoporousmaterials, have drawn considerable attention as promising adsorbentsfor gas separations. Among various separation applications, CO2/CO separation is of particular interest owing to its industrialrelevance. While searching for promising MOFs from tens of thousandsof candidates represents a great challenge, this study conducts large-scalemolecular simulations to identify top-performing CO2 adsorbents,followed by investigating structure-property relationshipsfor their design. Optimal MOFs are found to possess features suchas metal nodes of greater metallic charges and dipole moments witha relatively confined pore structure. With the large-scale data atour disposal, machine learning models capable of predicting the CO2-to-CO selectivity and adsorption uptakes are also established.Specifically, three algorithms including support vector regression(SVR), extreme gradient boosting (XGBoost), and random forest (RF)models are employed. The results show that the RF algorithm demonstratesthe best accuracy, and the r value for the predictedCO(2)-to-CO selectivity (S) can be as largeas & SIM;0.88. The relative importance of the adopted features isalso investigated with results suggesting that the adsorption of CO2 initiates more preferentially than that of CO due to thestronger van der Waals interaction and electrostatic contributionbetween CO2 and the metal sites. Finally, a design ruleis proposed for the optimal design of CO2-selective materials.Overall, this work demonstrates a successful hybrid approach combiningmolecular simulations and machine learning for screening highly CO2/CO selective MOFs and offering insights into the design ofoptimal adsorbents.