Multiscale analysis of breakdown strength in polyimide nanocomposite films: Simulations, experiments, and machine learning

Abstract

Polyimide inorganic nanocomposites have become a major focus of research in various fields due to their excellent integrated properties. This study adopts a combination of theoretical simulations and experimental methods. Using Density Functional Theory (DFT) calculations, the microscopic properties of polyimide molecules and nano-TiO₂ and nano-Al₂O₃ were investigated, while Molecular Dynamics (MD) simulations revealed the microscopic mechanism behind the thermal breakdown of polyimide films. Subsequently, polyimide/titanium dioxide (PI/nano-TiO₂) and polyimide/titanium dioxide/alumina (PI/nano-TiO₂/nano-Al₂O₃) nanocomposite films were prepared. The results showed that the breakdown strength of the PI/nano-TiO₂ film was 3.58 times that of conventional polyimide films. The surface modification of nanoparticles was achieved by adding the coupling agent KH550, which enhanced the structure, morphology, and overall properties of the composite films, significantly improving their thermal stability. Our research framework adopts a multi-scale approach, starting with insights at the microscopic level obtained from Density Functional Theory (DFT) calculations and MD simulations, followed by experimental validation, and ultimately classifying the breakdown strength through machine learning. We systematically analyzed the breakdown performance of polyimide films across multiple scales, from molecular simulations and experimental analysis to machine learning, providing theoretical support and practical reference for the design of nanocomposite films, and demonstrating the potential of this material for high-performance dielectric applications.