Artificial intelligence simulations for critical variable tracking during polymer synthesis

The success of experimental studies relies heavily on human factors, and the quality and amount of resulting data is inherently limited. To overcome these limitations, big data-based artificial intelligence (AI) methods are used to rapidly learn complex theories and analyze experimental data to predict new outcomes. Living anionic ring-opening polymerization (LAROP) is a valuable method for synthesizing polymers; however, the stringent processing conditions (low humidity and oxygen contents) and multiple variables make it difficult to identify the most influential factor. This study employed a group method of data handling a polynomial neural network as an AI method to predict critical synthesis variables during the LAROP synthesis of poly(ethylene oxide-b-allyl glycidyl ether) (PEO–PAGE) as a representative block copolymer. Owing to the O-groups at the end of PEO blocks, monomers in the glycidyl ether group propagate at the end of these blocks. PAGE blocks can propagate up to a mass fraction ratio of less than 1:1.5 (PEO:PAGE blocks). This study provides insights into the LAROP synthesis of PAGE blocks with long hydrophobic chains, offering valuable information for advancing polymer synthesis techniques. Furthermore, the findings support the use of AI methods for predicting experimental synthesis conditions to improve the efficiency of polymer development.