(51g) Generating, Simulating, and Synthesizing Novel Materials for Gas Storage & Separations

Metal-organic frameworks (MOFs), porous-aromatic frameworks (PAFs), porous-organic polymers (POPs) and other material classes each contain millions of candidate materials. While computational screening methods reported to date have significantly accelerated the identification of promising porous materials relative to experimental trial-and-error, they are not matched to the enormous scale of porous materials. In this presentation, we describe an enumerative structure generating algorithm that can output millions of novel hypothetical materials per CPU per day using an input library of chemical building blocks. Importantly, each generated structure is associated with a de facto synthesis protocol, based on the widely employed self-assembly of rigid building blocks in solution. These structures can be screened for gas adsorption properties using grand canonical Monte Carlo (GCMC) simulations in parallel on high performance computers. For simulations of polar gas molecules, such as CO2, partial charges on the framework atoms must be estimated in a computationally efficient fashion. To address this problem, we developed a non-iterative scheme for estimating partial charges in periodic structures. Collectively these tools enable a step-change in the pace of materials discovery, and we have already released select software codes into the public domain.