Accelerating Energetic Ionic Liquid Discovery: A Synergistic Fusion Strategy for Property Prediction

Energetic ionic liquids (EILs) represent a promising class of energetic materials, distinguished by their ultralow vapor pressure, reduced sensitivity, and highly tunable molecular architectures. However, their development remains largely dependent on empirical trial-and-error approaches, posing significant challenges for accurate a priori prediction of key performance metrics across unexplored chemical spaces. To overcome these limitations, we developed a predictive framework based on a combined feature and model fusion strategy. A comprehensive EIL database was curated from the literature, incorporating a broad range of electronic, topological, and thermodynamic descriptors. These features were integrated through a synergistic fusion strategy, and multiple machine learning models were ensembled to enable accurate prediction of eight critical properties: density, melting point, decomposition temperature, specific impulse, glass transition temperature, vacuum-specific impulse, ignition delay time, and heat of formation. Among these, models for decomposition temperature and density demonstrated excellent generalizability, achieving mean absolute errors below 22.2 °C and 0.032 g·cm^–3, respectively. This framework bridges molecular-level descriptors with macroscopic performance, offering a scalable and data-driven alternative to experimental screening. By enabling multiproperty prediction with improved accuracy and efficiency, our approach provides a powerful tool for accelerating the design of novel EILs.