(651a) Automatic Rule-Based Mapping and Evaluation of the Synthesis Space of a Sustainable Small-Molecule Refinery

Molecules, such as N₂, H₂S, CO₂, and H₂O are abundant and freely available sources of C, H, S, O, and N. A number of emerging chemical and biological processes are being explored to sustainably upgrade these molecules to value-added products. In this context, it is valuable to map out what molecules can be made from these small-molecule feedstocks and how “best” to make them. While such an exercise offers the possibility of discovering new molecules and synthesis routes, the synthesis space of potential molecules is vast because a plethora of upgrading options exist. This precludes a manual analysis of the product space, the optimal product portfolio, and ideal synthesis routes.

In this talk, we use a scalable computational approach to explore the synthetically feasible product space obtained from small and abundant molecules. In particular, we use the reaction network generator (RING) to automatically construct a comprehensive synthesis network accessible from these feed molecules using established electrochemical and thermochemical reaction rules as input. We posit that the molecules in this network represents the synthetically feasible space. Subsequently, we characterize these molecules in terms of their physicochemical properties, whether they are present in molecule databases, the energetics of synthesis pathways, etc.