EMFF-2025: a general neural network potential for energetic materials with C, H, N, and O elements
Mingjie Wen, Jiahe Han, Wenjuan Li, Xiaoya Chang, Qingzhao Chu, Dongping Chen
Abstract
The discovery and optimization of high-energy materials (HEMs) face challenges due to the computational expense and slow iteration of traditional methods. Neural network potentials (NNPs) have emerged as an efficient alternative to first-principles simulations. This study presents EMFF-2025, a general NNP model for C, H, N, and O-based HEMs, leveraging transfer learning with minimal data from DFT calculations. The model achieves DFT-level accuracy, predicting the structure, mechanical properties, and decomposition characteristics of 20 HEMs. Integrating EMFF-2025 with PCA and correlation heatmaps, we map the chemical space and structural evolution of these HEMs across temperatures. Surprisingly, EMFF-2025 uncovers that most HEMs follow similar high-temperature decomposition mechanisms, challenging the conventional view of material-specific behavior. EMFF-2025 offers a versatile computational framework for accelerating HEM design and optimization.