Illuminating the Interface of Blocc Chemistry and Data Science: Maximizing Function with ML-Guided Discovery and a...

The intersection of automatable blocc chemistry for iterative carbon-carbon bond formation with artificial intelligence is amplifying molecular innovation in new and exciting ways. In this lab, students are introduced to concepts and tools that help them gain familiarity and confidence with this emerging area of chemistry. Students specifically learn about four automated synthesis platforms, each of which stitches together a bounded set of molecular building blocks using just one type of robust bond-forming reaction. Students then analyze a variety of small molecules and biopolymers to identify the most redundant types of bonds in each molecule that are compatible with iterative formation from bifunctional building blocks. Based on their analysis, students then select an appropriate iterative automated synthesis platform and determine which molecular building blocks would be required to assemble the desired targets. Students next investigate two case studies where blocc chemistry was used used in concert with artificial intelligence to discover new molecular functions. In the first case, students identify high-performing blocks from selected data sets to optimize a single objective function related to organic laser properties. In the second case, students engage with a new online platform inspired by Scratch, dubbed the Digital Molecule Maker (DMM), to perform a multi-objective optimization of organic photovoltaic candidates (OPV). To conclude the lab, students manually perform the blocc chemistry that is foundational for the DMM. This activity series is the second installment of a sequence of undergraduate laboratories designed to illuminate the functional discovery-enabling interface of AI and automated blocc chemistry.