General quantitative structure–property relationships and machine learning correlations to energetic material sensitivities

Which molecular properties determine the impact sensitivity of an explosive? A machine learning quantitative investigation of nitroaromatic explosives

“…Additionally, this plethora of possible influences shows that the prediction of impact sensitivity based on physical properties “remains a challenge”. 26…”

“…, dissociation energies), among many other properties, known as descriptors in QSPR jargon. 19,26,36,37 Quantum chemical modeling of impact sensitivities, in particular, has quite been quite successful in this regard, especially for a given class of materials such as nitroaromatics, nitramines, among others. Representative works of this approach include correlating impact sensitivities with molecular properties such as bond strengths and molecular electrostatic potentials, 30,38–40 whereas others employed Mulliken charge values of nitro groups 15,41 and examined binding forces using the Wiberg bonding index.…”