(167t) Prediction of Carbon-Dioxide Sorption in Polymer/Ionic-Liquids Systems

The demand for energy has been growing due to economic, technological growth, and human activities. This growing demand and the reliance on fossil fuels, as the major energy resource, lead to an increase in the emission of greenhouse gases. Current methods of CO₂ capture rely on amine solutions where they are challenges associated with amine regeneration, toxicity, and high energy cost. To mitigate these challenges, many efforts are ongoing to utilize new materials for CO₂ capture. Among the new materials and technologies, Ionic liquids (IL) are promising candidates for CO₂ capture and storage applications because of their unique properties such as thermal stability, nontoxicity, and low volatility. Nevertheless, IL’s usage has been hindered by their high viscosity. This problem can be circumvented by: changing the separation scheme, making absorbents for pressure swing processes, increasing the fluoroalkyl chain in the anion in order to reduce the high viscosity of IL. To formulate absorbents, IL can be immobilized within a polymers matrix. Recent studies demonstrated the role of interfacial phenomena and IL ordering and its influence on gas sorption in IL mixture in polymers. As Machine Learning (ML) has become a powerful tool due to its potential of assisting in the polymeric field, this tool is utilized in carbon capture applications. In this study, we investigated the effect of Graphene Nanoplatelets (GNPs) on the macrostructures of the casted film by probing the variations in gas sorption capacities from these mixtures. We also developed a ML model to predict CO₂ sorption within polymeric mixtures and discuss the potential of applying ML to the gas sorption prediction from the molecular to process level. By using the 3-layer Neural Network model, we successfully achieved a good agreement between experimental data and model predictions for all Polymer/IL mixtures with an overall R² value of 92%.