Single-Image-Based Deep Learning for Precise Atomic Defect Identification.

Defect engineering is widely used to impart the desired functionalities on materials. Despite the widespread application of atomic-resolution scanning transmission electron microscopy (STEM), traditional methods for defect analysis are highly sensitive to random noise and human bias. While deep learning (DL) presents a viable alternative, it requires extensive amounts of training data with labeled ground truth. Herein, employing cycle generative adversarial networks (CycleGAN) and U-Nets, we propose a method based on a single experimental STEM image to tackle high annotation costs and image noise for defect detection. Not only atomic defects but also oxygen dopants in monolayer MoS2 are visualized. The method can be readily extended to other two-dimensional systems, as the training is based on unit-cell-level images. Therefore, our results outline novel ways to train the model with minimal data sets, offering great opportunities to fully exploit the power of DL in the materials science community.

Chiral Stacking Identification of Two-Dimensional Triclinic Crystals Enabled by Machine Learning.

Chiral materials possess broken inversion and mirror symmetry and show great potential in the application of next-generation optic, electronic, and spintronic devices. Two-dimensional (2D) chiral crystals have planar chirality, which is nonsuperimposable on their 2D enantiomers by any rotation about the axis perpendicular to the substrate. The degree of freedom to construct vertical stacking of 2D monolayer enantiomers offers the possibility of chiral manipulation for designed properties by creating multilayers with either a racemic or enantiomerically pure stacking order. However, the rapid recognition of the relative proportion of two enantiomers becomes demanding due to the complexity of stacking orders of 2D chiral crystals. Here, we report the unambiguous identification of racemic and enantiomerically pure stackings for layered ReSe2 and ReS2 using circular polarized Raman spectroscopy. The chiral Raman response is successfully manipulated by the enantiomer proportion, and the stacking orders of multilayer ReSe2 and ReS2 can be completely clarified with the help of second harmonic generation and scanning transmission electron microscopy measurements. Finally, we trained an artificial intelligent Spectra Classification Assistant to predict the chirality and the complete crystallographic structures of multilayer ReSe2 from a single circular polarized Raman spectrum with the accuracy reaching 0.9417 ± 0.0059.