Discovery of liquid crystalline polymers with high thermal conductivity using machine learning

Hayato Maeda; Stephen Wu; Rika Marui; Erina Yoshida; Kan Hatakeyama-Sato; Yuta Nabae; Shiori Nakagawa; Meguya Ryu; Ryohei Ishige; Yoh Noguchi; Yoshihiro Hayashi; Masashi Ishii; Isao Kuwajima; Felix Jiang; Xuan Thang Vu; Sven Ingebrandt; Masatoshi Tokita; Junko Morikawa; Ryo Yoshida; Teruaki Hayakawa

doi:10.1038/s41524-025-01671-w

Article Discovery of liquid crystalline polymers with high thermal conductivity using machine learning

Hayato Maeda ; Stephen Wu ; Rika Marui ; Erina Yoshida ; Kan Hatakeyama-Sato ; Yuta Nabae ; Shiori Nakagawa ; Meguya Ryu ; Ryohei Ishige ; Yoh Noguchi ; Yoshihiro Hayashi ; Masashi Ishii ; Isao Kuwajima ; Felix Jiang ; Xuan Thang Vu ; Sven Ingebrandt ; Masatoshi Tokita ; Junko Morikawa ; Ryo Yoshida ; Teruaki Hayakawa

Download file (3.05 MB)
Download as zip (2.54 MB)

Collection

Citation

Hayato Maeda, Stephen Wu, Rika Marui, Erina Yoshida, Kan Hatakeyama-Sato, Yuta Nabae, Shiori Nakagawa, Meguya Ryu, Ryohei Ishige, Yoh Noguchi, Yoshihiro Hayashi, Masashi Ishii, Isao Kuwajima, Felix Jiang, Xuan Thang Vu, Sven Ingebrandt, Masatoshi Tokita, Junko Morikawa, Ryo Yoshida, Teruaki Hayakawa. Discovery of liquid crystalline polymers with high thermal conductivity using machine learning. npj Computational Materials. 2025, 11 (1), 205. https://doi.org/10.1038/s41524-025-01671-w

(BibTeX)

Description:

(abstract)

It is highly desired to enhance the heat conduction of polymeric heat-dissipating materials. However, the thermal conductivity of polymeric materials is one to three orders of magnitude lower than that of metals or ceramics due to the occurrence of phonon scattering at amorphous portions. Several attempts have been made to overcome this empirical or theoretical limit to increase the heat flow in the orientation axis by liquid crystallizing the polymers. However, the design of polymeric liquid crystals remains largely empirical. In this study, we have developed a machine learning model that can predict with more than 96% accuracy whether or not assembled polymers form a liquid crystalline state with compositional or structural features of any given polymer repeating unit. Exploring the inverse mapping of the model, we identified a comprehensive set of chemical structures for liquid crystalline polyimides. All the synthesized polymers were experimentally confirmed to spontaneously form liquid crystalline structures with their thermal conductivities ranging from 0.722 to 1.26 W/(m · K).

Rights: