Quantum and Classical Machine Learning Studies of Optimal III–Nitride Epitaxy

Considerable interest exists in the development of machine learning–based approaches for identifying optimal materials designs and synthesis conditions. In this work, data describing over 100 plasma–assisted molecular beam epitaxy (PAMBE) growth trials each of GaN and InN have been organized into separate, composition–specific data sets. For each growth record, the complete set of experiment parameters are associated with a measure of crystallinity as determined by reflection high–energy electron diffraction (RHEED) patterns. Additionally, a Brag–Williams measure of lattice disorder (S²) is included as a second figure of merit for investigation. Quantum and classical supervised learning algorithms – including logistic regression, tree–based algorithms, and quantum variational circuits – are trained on the data and used to study which growth parameters are most statistically important for influencing crystallinity and S². The probabilities of obtaining monocrystalline GaN and InN thin film crystals are predicted across processing spaces of the two most important synthesis parameters. The machine learning predictions of these growth characteristics agree with published values for obtaining single crystalline GaN and InN thin films. S² is also predicted across the same processing spaces. The predictions indicate that different growth conditions are of interest depending on whether a single crystalline sample or a well–ordered lattice (as measured by S²) is desired.