Heuristic Search Strategy for Transforming Microstructural Patterns to Optimal Copolymerization Recipes

Abstract

Manipulation and optimization of copolymer microstructure for tailoring final properties is of great importance in macromolecular science and engineering. Uncovering the complexities of the interrelationships between copolymerization recipe and copolymer microstructure (a challenging field of study in its own right) is a multi-objective optimization problem, which has attracted a lot of attention in the last 10-15 years. In the present study, a powerful optimizer was developed based on the Non-dominated Sorting Genetic Algorithm (NSGA-II) for transforming desired microstructural copolymerization profiles, including molecular weight distribution (MWD) and chemical composition distribution (CCD), back to optimal copolymerization recipes and operating conditions. The optimizer developed has the beneficial features of robust machine learning and multi-objective optimization based upon heuristic search strategies. The metallocene-catalyzed ethylene/α-olefin copolymerization was selected as a sufficiently complex system to challenge the proposed optimization tool. The developed computer code was used to explore copolymerization recipes (polymerization temperature and concentrations of ethylene, 1­butene, cocatalyst, and hydrogen) needed to synthesize copolymers having desired microstructural features. Based on the results obtained, it is now possible to produce various grades or tailor-make the copolymer structure by suggesting the ‘best’ copolymerization recipe/conditions as reliably as possible.