A lamellar-morphology-based computational modeling for predicting the thermal conductivity of semicrystalline polymers

Molecular-scale design of crystal structures is an emerging approach for significantly improving the intrinsic properties of polymers. In this study, we performed computational modeling to quantitatively predict the thermal conductivity of polymers based on their crystal morphology. Polyethylene lamellae with alternating crystalline and amorphous phases were prepared using isothermal crystallization within a coarse-grained molecular dynamics framework. The crystalline bulk, transient, and amorphous regions were clearly distinguished based on the distribution of the Steinhardt-bond order parameter. The thermal conductivity of the local regions was estimated discretely by applying a steady heat flow, facilitating a bottom-up prediction of the thermal response depending on the morphology of the lamellae. In particular, the integration with homogenization theory can evaluate the effective thermal conductivity of the system in terms of crystallinity, temperature, lamella thickness and orientation. The modeling successfully provided quantitative predictions for two representative hierarchical structures: spherulites and oriented crystals. Therefore, this study serves as a theoretical guide for molecular-level rational design of semicrystalline materials with high thermal conductivity.