Putting an end to the handcuffs of the periodic table: Using atomic features for predicting detonation velocity

The traditional detonation models are limited by their dependence on elemental composition as the primary feature. This means that specific elements must be used as inputs, resulting in a tedious process of tuning the models for every new element that is introduced. This approach hinders the scalability and adaptability of the models, making them less efficient in accommodating the vast diversity of materials encountered in energetic chemistry. The constant need for adjustment and calibration for each element poses a significant obstacle in terms of time and resources, which ultimately impacts the models' ability to provide accurate predictions for a broad spectrum of compounds. For instance, an extension of Kamlet and Jacobs by Pang et al. had to be developed to account for metals. This required to use of data obtained from EXPLO5 using the BKW EOS, due to a lack of experimental values of transition metal energetics, meaning it is a model built on top of a model.This approach is acceptable, if it is for initial validation, however, the field should not rely on such models for predicting novel materials. As the field of energetic materials advances, and the limits of CHNO explosives have been pushed, it has become apparent that more generalized and flexible models are crucial to overcome the limitations associated with the elemental composition-centric paradigm. In this study, a Gaussian process regressor is used to predict the detonation velocity. As features atomic values were used, which allows the application of the model to explosives of any elemental compositions. This approach does not require reparameterization for every element and can provide useful estimations for the detonation velocity of explosives. For initial validation purposes, this study will interpolate from materials containing elements from the first and second periods to explosives-containing materials in the third period.