Light responsive polyimides with liquid crystal (LC) alignment capability are often used in display devices. Meanwhile, a typical polyimide exhibits a low thermal conductivity, which is not suitable for applications in heat-prone flexible electronics. Recent research has reported successful fabrication of high intrinsic thermal conductivity LC polyimide films that the in-plane thermal conductivity exceeds 2 W/(mK). In this study, we exploited the rapidly increasing amount of polymer data driven by the new interdisciplinary research field, called polymer informatics, to accelerate the discovery of LC polyimide with desired thermal properties. We began with a newly designed template for building the backbone of polyimides. Purchasable molecules from the ZINC database that match the components in our template were exhaustively searched and extracted to construct a large virtual library of polyimides. Then, a machine learning model for LC classification was trained with data from PoLyInfo, a polymer property database, and used to screen out promising candidates of LC polyimide. Thermal conductivity of the screened candidates was predicted by another machine learning model trained with data from molecular dynamics simulations. Finally, guided by the predicted thermal properties of the candidates, we successfully synthesized new LC polyimides, which demonstrated the effectiveness of our design approach.

Data-driven design and synthesis of new liquid crystal polyimides

Fall 2022 abstracts, posters, and presentations will launch in mid-September.

ACS Fall 2022

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.