The self-driving chemist aims to develop an autonomous reaction cycle to drive discovery through chemical synthesis.  It can generate and test hypotheses with relatively few actions required by humans. It allows chemical researchers to divert their attention away from routine experimental tasks and focus on complex problem solving aided by machine learning and artificial intelligence.

Argonne’s self-driving chemist platform in the Center for Nanoscale Materials (CNM) works on the basis of a learned general association between molecular structures and reactivity, as well as machine learning methods that can process data and assess reactivity without human intervention. Components include a controlled experimental environment, liquid and powder handling systems, reactors arrays, and filtration chemical storage systems, and characterization modules.

In one project, we are applying the self-driving chemist to determine deconstruction pathways for polymer electronics. The plastic waste they generate pose severe environmental problems. The self-driving chemist is also being used for discovering the mechanisms for cross-coupling reactions in organic chemistry, where two fragments are joined together with the aid of a metal catalyst. The selectivity and efficiency of these cross-coupling reactions have been a long-term challenge due to their complex reaction components and conditions.

The self-driving chemist is part of Polybot, an autonomous discovery laboratory in CNM. Polybot includes self-driving materials synthesis and fabrication, as well as robotic sample transfer, experimentation, characterization, and data analysis.