The Use of Genetic Algorithms to Predict the Crystal Structures of Energetics

Crystallography data of high energy density compounds such as TATB, TNT, HMX, and CL-20 are difficult to obtain experimentally; experimental crystallography techniques are expensive and hold many safety risks. The complex energy landscapes of these large energetic compounds make it difficult and time-intensive to predict stable crystal structures via computational means. Genetic search algorithms are regularly used in crystal structure prediction of other compounds by reducing dimensionality of the landscape to minimize computing costs and avoid local minima. We will present the use of an established genetic algorithm for the application of predicting the crystal structures of TATB, TNT, HMX, and CL-20. The results will be validated through comparison of experimental densities, lattice parameters, and x-ray diffraction data. This work establishes a methodology for predicting crystal structures of new energetics and novel energetic co-crystals.