A research team from the East China University of Science and Technology has developed an innovative AI-assisted materials-genome approach to expedite the design of high-performance thermosetting polyimides. Their findings, published online on September 2, 2025, in the Chinese Journal of Polymer Science, aim to overcome long-standing challenges in optimizing the mechanical properties of polyimide films, which are crucial in various industries including aerospace and flexible electronics.
Polyimide films are renowned for their exceptional thermal stability and insulation properties. Yet, achieving a balance between high modulus and toughness in these materials has proven difficult. Previous methods of trial-and-error synthesis are often slow and costly, limiting the exploration of complex molecular structures. The new approach harnesses the power of artificial intelligence to learn structure-property relationships directly from extensive datasets.
The research introduces a machine-learning model designed to predict three essential mechanical parameters: Young’s modulus, tensile strength, and elongation at break. This model was trained on over 120 experimental datasets of polyimide films, allowing the team to analyze a vast chemical space of 1,720 phenylethynyl-terminated polyimides (PPIs). The outcome was the identification of a new formulation called PPI-TB, which exceeded the performance of established benchmark polyimides.
Innovative Methodology Validates Mechanical Performance
The research team employed Gaussian process regression (GPR) models to achieve high predictive accuracy, with R² values ranging from approximately 0.70 to 0.74 for the three mechanical metrics. These models were instrumental in scoring every candidate structure based on comprehensive mechanical performance. Subsequent molecular dynamics simulations validated the screening process, demonstrating that PPI-TB (gene combination A4/B32) exhibited superior modulus at 3.48 GPa, while also showing enhanced toughness and strength compared to existing polyimides such as PETI-1 and O-O-3.
The alignment between predicted and measured data was further confirmed through experiments conducted on representative PPIs. Detailed analyses of the “genes” involved highlighted critical design principles. For instance, the inclusion of conjugated aromatic structures was found to enhance stiffness, while the presence of heteroatoms and heterocycles improved molecular interactions. Additionally, flexible silicon- or sulfur-containing units contributed to increased elongation.
Transforming Polymer Design with AI Insights
According to Prof. Li-Quan Wang, one of the study’s corresponding authors, the integration of AI into polymer design is akin to decoding a genome. “Machine learning not only predicts performance but also reveals which chemical ‘genes’ are driving it,” he stated. This synergy between data science and chemistry enables the exploration of material possibilities that would traditionally require decades of research.
The materials-genome strategy developed by the team offers a universal and scalable framework for creating polymers with tailored combinations of stiffness, strength, and flexibility. These attributes are vital for applications in microelectronics, aerospace composites, and flexible circuit substrates. By replacing years of experimental iteration with predictive modeling and virtual screening, this innovative method significantly reduces both costs and development time.
Beyond the realm of polyimides, the workflow promises to be adaptable to other high-performance polymer classes. This adaptability could guide the creation of lightweight, durable, and thermally stable materials that are pivotal for future advancements in electronic and aerospace technologies.
The research received financial support from the National Key R&D Program of China and the National Natural Science Foundation of China. The implications of this work underscore the transformative potential of AI in materials science, paving the way for the next generation of high-temperature polymers.
