Machine learning-assisted systematical polymerization planning: case studies on reversible-deactivation radical polymerization

The combined influence of chemical composition, molecular weight (MW) and molecular weight distribution (D) on the functions and performances of polymeric materials necessitates simultaneous satisfaction of multidimensional requirements during polymer synthesis. However, the complexity of polymerization reactions often dissuades chemists when precisely accessing diversified polymer targets. Herein, we developed a machine learning (ML)-assisted systematical polymerization planning (SPP) platform for addressing this challenge. With ML model providing integrated navigation of the reaction space, this approach can conduct multivariate analysis to uncover complex interactions between the polymerization result and conditions, prescribing optimal reaction conditions to achieve discretionary polymer targets concerning three dimensions including chemical composition, MW and D values. Given the increasing importance of polymerization in advanced material engineering, this ML-assisted SPP platform provides a universal strategy to access tailored polymers with on-demand prediction of polymerization parameters.

Automated summary

The study developed a machine learning-assisted systematic polymerization planning platform that can meet the needs in multiple dimensions such as chemical composition, molecular weight, and molecular weight distribution, and provide optimal reaction conditions to achieve customized polymer targets.