(346ar) Evaluating the Adsorption Landscape for Polymers of Intrinsic Microporosity

The potential for microporous materials to provide energy-efficient solutions for chemical separations or storage has provided motivation for the development of many families of porous systems; for example, metal organic frameworks, covalent organic frameworks, zeolites, and microporous polymers. Many of these microporous materials are modular in nature or can be readily modified, which yields a substantial set of conceivable adsorbents. Studying these sets at variable operating conditions or for a diversity of adsorbates can be thought of as evaluating adsorption space. Molecular simulation screening studies provide a useful tool for this type of challenge because of their capability to efficiently predict adsorption properties for a large batch of systems, which can ultimately expedite adsorbent design/discovery. The presented study has applied a simulated screening approach for 15 unique polymers of intrinsic microporosity (PIMs) with 20+ distinct adsorbate species. To produce the full adsorption isotherms, a combination of molecular dynamics (MD) and Monte Carlo (MC) techniques are utilized, which greatly improves the accuracy of predictions made in the presence of appreciable adsorbate uptake by accounting for sorption-induced polymer chain rearrangement. This adsorption space screening study that incorporates polymer chain dynamics is the first of its kind and has become possible because of our newly developed python interface between RASPA and LAMMPS, which handle the Monte Carlo and Molecular Dynamics simulations, respectively. Additionally, to efficiently manage computational resources we have included a set of automated convergence statistics that allows the MC/MD process to terminate upon reaching the equilibrated state used for predictions, thus minimizing excess wall time. Using these methods, we have determined the onset swelling pressures for all of the adsorbate-adsorbent pairs, which is a quantity of industrial relevance. Finally, in an effort to circumvent laborious simulations, we examine the applicability of developing a model for the rapid prediction of adsorption properties in PIMs based on commonly calculated values, such as Henry’s constant and helium void fraction.