Active Learning as a Tool for Optimizing “Plug‐n‐Play” Electrochemical Atom Transfer Radical Polymerization

Abstract

A recently reported “plug‐n‐play” approach to simplified electrochemical atom transfer radical polymerization (seATRP) is investigated using machine learning. It is shown that Bayesian optimization via an active learning (AL) algorithm accelerates optimization of the polymerization of oligo(ethylene glycol methyl ether acrylate)480 (OEGA480) in water. Molecular weight distribution (Mw/Mn; dispersity; Ɖm) is the output selected for optimization targeting poly(oligo[ethylene glycol methyl ether acrylate]) (POEGA480) with low dispersity (Ɖm < 1.30). Input variables included applied potential (Eapp), [M] and [M]/[I], which led to a potential space of 275 possible reaction conditions. From a training data set of seven reactions, selected to yield uncontrolled POEGA480 with higher dispersities (Ɖm > 1.5), ten iteration loops are performed. During each iteration the algorithm suggests the next reaction conditions. The reactions are then performed and the conversion, number average molecular weight (Mn) and Ɖm values are recorded and the Ɖm values fed back into the algorithm. Overall, 80% of the experiments yield POEGA with Ɖm < 1.30. Conversely, only 30% of experiments performed using reaction conditions selected at random from the possible reaction space yield POEGA with Ɖm < 1.30. This study suggests that adopting AL methods can improve the efficiency of optimizing a given seATRP reaction.

Bayesian optimization of radical polymerization reactions in a flow synthesis system

“…[42][43][44][45][46][47] Flow synthesis offers several benefits, including homogeneous mixing, efficient heat exchange, precise residence time control and scalability [48][49][50] , making it a suitable candidate for integration with machine learning techniques. [51][52][53][54][55][56][57] Warren et al previously developed a polymer synthesis and analysis platform that allowed RAFT polymerization to be automated and used this system to explore the trade-off between molar mass dispersity and monomer conversion 58 . In our previous work, we developed a flow copolymerization system based on a microflow mixer and explored the relationship between process conditions and reactivity ratios in styrene -methyl methacrylate (St-MMA) and glycidyl methacrylate -MMA copolymers.…”

Bayesian optimization of radical polymerization reactions in a flow synthesis system

“…[42][43][44][45][46][47] Flow synthesis offers several benefits, including homogeneous mixing, efficient heat exchange, precise residence time control and scalability [48][49][50] , making it a suitable candidate for integration with machine learning techniques. [51][52][53][54][55][56][57] Warren et al previously developed a polymer synthesis and analysis platform that allowed RAFT polymerization to be automated and used this system to explore the trade-off between molar mass dispersity and monomer conversion 58 . In our previous work, we developed a flow copolymerization system based on a microflow mixer and explored the relationship between process conditions and reactivity ratios in styrene -methyl methacrylate (St-MMA) and glycidyl methacrylate -MMA copolymers.…”

An electrochemical Hofmann rearrangement on acrylamide copolymers

“…Under these conditions aryl and alkyl amides are converted to O-methyl carbamates in good yields. 33 Herein, we build on our recent work on the electrochemically triggered/mediated synthesis of polymers (by eATRP) [34][35][36][37] by adopting the eHofmann rearrangement to achieve mild and efficient modification of acrylamide containing copolymer scaffolds for the first time. Overall we demonstrate that the primary amide side-chains can serve as 'masked' isocyanates which can be captured by alcohol-based nucleophiles forming O-alkyl carbamate side chains in the process.…”