One of the most important challenges in chemistry research is designing and discovering molecules with useful properties because this requires exploring extremely large chemical spaces. Self-driving labs have been proposed as a solution. Integrating property prediction, synthesis, analysis, characterization and experiment planning, they can close the chemical discovery “loop” and make faster, more efficient progress. Automated synthesis platforms (ASPs) are a key component of self-driving labs, but they are currently limited by the scope of reactions that they can carry out, while humans can perform a much greater variety of reactions. Understanding what areas of chemical space ASPs can and cannot reach is important for effectively utilizing self-driving labs. We present a framework for evaluating synthetic routes that combine human and automated synthesis like a subway map (a). Humans carry out slower, more complex reactions (blue, a & b), just like walking to or from a subway station. ASPs perform a limited set of reactions that rapidly traverse chemical space to limited, pre-determined areas (pink, a & b), like riding the subway. With this approach, self-driving labs could reach previously inaccessible molecules from cheap starting materials. Our approach evaluates a synthetic route in terms of cost in time, money and materials per distance traveled through chemical space. As a demonstration, we simulate a self-driving lab carrying out a molecular design and property optimization campaign on a space of ~3500 molecules.

Punching the ticket to more efficient materials development

ACS Spring Meeting 2021

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.