EXTENDED ABSTRACT: Advanced polymeric materials play an important role in the ffelds of aerospace and civilian. However, the traditional trial-and-error method makes it difffcult to address some long-standing material design challenges. Research paradigms have evolved from experimental, theoretical, and computational simulation to the 4th paradigm based on artiffcial intelligence (AI) [1]. Applying Al to the material research is challenging. The presenter and his team focused on polymer materials by collecting data, developing digital methods, building databases, developing machine learning (ML) methods suitable for polymers, and establishing AI performance prediction models. They developed a new structural design based on ML and material genomes to design and prepare a series of novel polymer materials [2-4]. For example, they established a database of epoxy resins, developed descriptors to describe the crosslinking structure, and built ML prediction models of epoxy thermoset properties. Based on this, they designed epoxy resins with excellent mechanical properties, and the mechanical properties of their composites were greatly improved [5]. In addition, they designed hybrid resins with high heat resistance and good processibility (as shown in Figure 1) and also with high toughness [4, 6].

Keywords: machine learning; material genome; high-performance resin; structural design

REFERENCES:

[1] L. Gao, J. Lin, L. Wang, L. Du, Acc. Mater. Res. 5, (2024), 571-584

[2] X. Xu, W. Zhao, L. Wang, J. Lin, L. Du, Chem. Sci. 14, (2023), 10203-10211

[3] X. Xu, W. Zhao, Y. Hu, L. Wang, J. Lin, H. Qi, L. Du. J. Mater. Chem. A 11, (2023), 12918-12927

[4] S. Zhang, S. Du, L. Wang, J. Lin, L. Du, X. Xu, L. Gao, Chem. Eng. J. 448, (2022), 137643

[5] Y. Hu, W. Zhao, L. Wang, J. Lin, L. Du, ACS Appl. Mater. Interfaces 14, (2022), 55004-55016

[6] J. Zhu, M. Chu, Z. Chen, L. Wang, J. Lin, L. Du, Chem. Mater. 32, (2020), 4527-4535