Single-atom level symmetry breaking visualized with STEM

The study of phase transitions in a material is fundamental to the understanding of its properties and functionalities. This pursuit is pivotal in the quest to decipher fundamental mechanisms of emergent phenomena such as superconductivity and quantum spin-liquid behaviors, many of which hold the key to enabling next-gen computing and energy technologies.

Current experimental techniques look mostly at macroscopically averaged parameters and do not provide information on spatially non-uniform states and local fluctuations. While there exist methods that combine experimental data with theoretical modeling to study these transitions with more details, there lacks an experimental method to verify these predictions.

Vasudevan et al. present a way to measure structural phase transitions at the atomic level by using machine learning analysis to interpret data from scanning transmission electron microscopy (STEM), establishing structure-property relationships on the atomic level.

They tested their method by looking at the transition from a trigonal prismatic to a distorted octahedra structure in the layered chalcogenide alloy MoS₂-ReS₂, as a function of the material’s chemical composition. They identified the parameters that can be used as descriptors for symmetry breaking on the atomic level, which can help understand the local mechanisms in doping-induced symmetry breaking.

Although the authors have only tested the method on a 2D material, they pose that this approach should also work on 3D materials in principle, as long as the structural distortions are isotropic or periodic along the beam direction.

“This method can also be extended to answer the fundamental causal questions in the physics of complex materials, for example, whether it is the local chemistry or global electronic concentration that control phase transition,” said author Sergei Kalinin.

Source: “Investigating phase transitions from local crystallographic analysis based on statistical learning of atomic environments in 2D MoS₂-ReS₂,” by Rama K. Vasudevan, Maxim Ziatdinov, Vinit Sharma, Mark P. Oxley, Lukas Vlcek, Anna N. Morozovska, Eugene A. Eliseev, Shi-Ze Yang, Yongji Gong, Pulickel Ajayan, Wu Zhou, Matthew F. Chisholm, and Sergei V. Kalinin, Applied Physics Reviews (2021). The article can be accessed at http://doi.org/10.1063/5.0012761 .