Search

Strain energy is a fundamental measure of the steric and configurational properties of organic molecules. The ability to estimate strain energy through quantum chemical simulations requires at minimum the knowledge of an initial set of nuclear coordinates. In general, such knowledge is not categorically known when screening or generating large numbers of molecule candidates in the context of molecular design. We present a machine learning approach to predict hydrocarbon strain energies using Benson group equivalents. A featurization strategy is crafted by concatenating the molecule group equivalent counts with easily computable molecular fingerprints. The data are obtained from electronic structure calculations we performed on a set of 166 previously synthesized strained hydrocarbons. These data are provided and include gas phase enthalpies of formation and associated optimized atomic coordinates. The strain energy prediction accuracy of several statistical learning methods is evaluated, and their respective merits and limitations are discussed.

Predictive models for the performance of explosives and propellants are important for their design, optimization, and safety. Thermochemical codes can predict some of these properties from fundamental quantities such as density and formation energies that can be obtained from first principles. Models that are simpler to evaluate are desirable for efficient, rapid screening of material screening. In addition, interpretable models can provide insight into the physics and chemistry of these materials that could be useful to direct new synthesis. Current state-of-the-art performance models are based on either the parametrization of physics-based expressions or data-driven approaches with minimal interpretability. We use parsimonious neural networks (PNNs) to discover interpretable models for the specific impulse of propellants and detonation velocity and pressure for explosives using data collected from the open literature. A combination of evolutionary optimization with custom neural networks explores and trains models with objective functions that balance accuracy and complexity. For all three properties of interest, we find interpretable models that are Pareto optimal in the accuracy and simplicity space.

For many experimentally measured chemical properties that cannot be directly computed from first-principles, the existing physics-based models do not extrapolate well to out-of-sample molecules, and experimental datasets themselves are too small for traditional machine learning (ML) approaches. To overcome these limitations, we apply a transfer learning approach, whereby we simultaneously train a multi-target regression model on a small number of molecules with experimentally measured values and a large number of molecules with related computed properties. We demonstrate this methodology on predicting the experimentally measured impact sensitivity of energetic crystals, finding that both characteristics of the computed dataset and model architecture are important to prediction accuracy of the small experimental dataset. Our directed-message passing neural network (D-MPNN) ML model using transfer learning outperforms direct-ML and physics-based models on a diverse test set, and the new methods described here are widely applicable to modeling many other structure-property relationships.