Decoding crystallography from high-resolution electron imaging and diffraction datasets with deep learning.

While machine learning has been making enormous strides in many technical areas, it is still massively underused in transmission electron microscopy. To address this, a convolutional neural network model was developed for reliable classification of crystal structures from small numbers of electron images and diffraction patterns with no preferred orientation. Diffraction data containing 571,340 individual crystals divided among seven families, 32 genera, and 230 space groups were used to train the network. Despite the highly imbalanced dataset, the network narrows down the space groups to the top two with over 70% confidence in the worst case and up to 95% in the common cases. As examples, we benchmarked against alloys to two-dimensional materials to cross-validate our deep-learning model against high-resolution transmission electron images and diffraction patterns. We present this result both as a research tool and deep-learning application for diffraction analysis.

Predictive multiphase evolution in Al-containing high-entropy alloys.

The ability to predict and understand phases in high-entropy alloys (HEAs) is still being debated, and primarily true predictive capabilities derive from the known thermodynamics of materials. The present work demonstrates that prior work using high-throughput first-principles calculations may be further utilized to provide direct insight into the temperature- and composition-dependent phase evolution in HEAs, particularly Al-containing HEAs with a strengthening multiphase microstructure. Using a simple model with parameters derived from first-principles calculations, we reproduce the major features associated with Al-containing phases, demonstrating a generalizable approach for exploring potential phase evolution where little experimental data exists. Neutron scattering, in situ microscopy, and calorimetry measurements suggest that our high-throughput Monte Carlo technique captures both qualitative and quantitative features for both intermetallic phase formation and microstructure evolution at lower temperatures. This study provides a simple approach to guide HEA development, including ordered multi-phase HEAs, which may prove valuable for structural applications.