High-accuracy measurement of the heat of detonation with good robustness by laser-induced breakdown spectroscopy of energetic materials

The heat of detonation of energetic materials (EMs) is determined by the release of chemical energy, bond energies, and chemical structures and can be reflected by the variation of emission intensities in laser-induced breakdown spectroscopy (LIBS). Herein, we propose a new method based on laser-induced breakdown spectroscopy, combined with small-sample machine learning, to accurately determine the heat of detonation by consuming small-dose samples. A statistical correction strategy is applied to improve the spectral quality and extract spectral features including the emission peak intensity and emission shape correlation intensity. Thereby, a high-accuracy quantitative model based on the plasma spectra is developed to predict the heat of detonation with RMSEC = 0.0314 kJ g⁻¹ and R_c² = 0.99. Excellent model robustness is verified through three independent tests at different dates, which exhibit a strong predictive power with RMSET′ = 0.1776, 0.1217, and 0.1207 kJ g⁻¹ and R_T′² = 0.98, 0.98, and 0.98, respectively. The elements of importance for analysis in the model further clarify that the quantitative diagnosis of the heat of detonation for EMs makes sense by LIBS. Therefore, this work can significantly facilitate the safe and fast determination of the heat of detonation of explosives in small-dosage samples.