Identification and Design of Better Diamine-Hardened\nEpoxy-Based Thermoset Shape Memory Polymers: Simulation and Machine\nLearning

An approach for designing thermoset\nshape memory polymers (TSMPs)\nwith improved shape memory properties through the integration of molecular\ndynamics (MD) simulation, machine learning (ML), and chemical intuition\nis presented. We identified key molecular features correlated with\ndesired shape memory properties, and used MD simulations to create\nan initial data set of TSMPs consisting of commercially available\nand manually designed monomers. Our prediction set was prepared by\nemploying four different approaches for modifying existing monomers\nbased on chemical intuition and insights gleaned from the literature.\nWe trained our ML model on the initial data set, used it to identify\nthe most promising candidates, evaluated their properties, and added\nthem to our initial data set. To further speed up the process, we\nidentified the most promising candidate after a few cycles and modified\nits structure to obtain a variant with better properties. Our approach,\nwhich capitalizes on the synergy between computational methodologies\nand human expertise to enable efficient exploration of vast chemical\nspace, resulted in the design of a monomer exhibiting more than 60%\nincrease in the desired recovery stress compared to the highest experimentally\nvalidated one.