(87b) Digital Chemistry and Engineering of Nanoporous Materials

Digital Chemistry, an emerging frontier in chemical science, integrates computational methods, data science, and automation to accelerate chemical research and applications. By leveraging automated workflows and data-driven insights, this approach optimizes important industrial processes such as drug discovery, materials design, and chemical synthesis, bridging theoretical innovation with practical implementation.

Nanoporous materials, including zeolites, metal-organic frameworks (MOFs), and porous organic polymers (POPs), hold transformative potential for tackling global challenges related to energy and the environment, such as hydrogen storage, carbon capture, and water harvesting. However, the extensive chemical and structural diversity of these materials presents significant challenges in identifying optimal candidates for specific applications.

In this presentation, I will demonstrate how our research group employs a Digital Chemistry framework, integrated with chemical engineering processes, to streamline the discovery, characterization, and evaluation of nanoporous materials. First, I will highlight the impact of the Computation-Ready, Experimental (CoRE) MOF database, demonstrating how open-access datasets enabled accurate prediction of high-surface area MOFs and accelerated the identification of high-performing materials for hydrocarbon separations.

Next, I will present a crystal graph convolutional network (GCN) model that we developed which could rapidly estimates DFT-derived partial atomic charges and band gaps in MOFs. This method has been applied to the newly updated CoRE MOF database, enabling large-scale accurate computational screening for carbon capture scenarios.

Finally, I will discuss the importance of multi-scale modeling in the real-world performance of nanoporous materials, particularly in complex multi-component gas separations. By coupling macrostate probability distributions (MPD) obtained from transition matrix Monte Carlo (TMMC) simulation with process modeling, we enhance both the efficiency and accuracy of mixture predictions.