Combining automated microfluidic experimentation with machine learning for efficient polymerization design

Polymerization reactions have traditionally been a notoriously difficult field of chemistry and chemical engineering due to the complexity of the systems, hazards associated with the reagents, environmental footprint of the operations and the highly non-linear topology of the reaction spaces. In this work, we aim to present a new paradigm for studying such complex reactions using machine-learning-assisted automated microchemical reactors. A custom-designed rapidly prototyped microreactor is used in conjunction with in situ infrared thermography and efficient, high-speed experimentation to map the reaction space for a metallocene polymerization catalyst. The study was able to reduce the volume of chemical waste generated by two orders of magnitude and provide the necessary data in an hour instead of months of traditional experimentation. Here we show that efficient microfluidic technology can be coupled with machine learning algorithms to obtain high-fidelity datasets on a complex chemical reaction.