Exploiting Machine Learning and Automated Synthesis in Continuous Flow for Process Optimization of the...

Area of Science:

• Polymer Chemistry
• Materials Science
• Chemical Engineering

Background:

• Polylactide (PLA) is a versatile, degradable, biobased polymer crucial for biomedical, packaging, and additive manufacturing.
• Optimizing PLA synthesis via ring-opening polymerization (ROP) is challenging, hindering its widespread adoption compared to petrochemical polymers.
• Organocatalyzed ROP of lactide offers a sustainable route, but requires precise control over reaction parameters.

Purpose of the Study:

• To systematically explore and optimize reaction conditions for the organocatalyzed ROP of l-lactide using a continuous flow reactor.
• To develop a predictive model for PLA synthesis kinetics and polymer characteristics based on experimental data.
• To identify optimal reaction parameters for maximizing PLA production rate and quality using multiobjective Pareto optimization.

Main Methods:

• Utilized a continuous flow reactor system for high-throughput experimentation of l-lactide ROP at room temperature.
• Employed 1,8-diaza-bicyclo[5.4.0]undec-7-ene (DBU) as catalyst and benzyl alcohol as initiator in dichloromethane solvent.
• Generated a comprehensive dataset analyzed with Kernel-Based Regularized Least Squares (KRLS) modeling and multiobjective Pareto optimization.

Main Results:

• A robust dataset capturing system kinetics and parameter dependencies was generated through high-throughput experimentation.
• KRLS modeling successfully predicted system kinetics and the influence of initial conditions on PLA characteristics.
• Pareto optimization identified conditions yielding high PLA conversion, excellent molecular weight control (low dispersity), and maximized production rates.

Conclusions:

• Continuous flow polymerization offers precise control over ROP kinetics for efficient PLA synthesis.
• Machine learning-assisted optimization significantly enhances the exploration of reaction spaces for polymer development.
• This study demonstrates a powerful, scalable approach for optimizing biobased polymer production and materials discovery.