1D domino-like phase transformation enables material programming in 2D MoTe ₂

Phase transformation is a fundamental phenomenon in nature, vital for both the scientific understanding and industrial applications of materials. The emergence of two-dimensional (2D) materials introduces new physical attributes that challenge traditional phase transformation theories due to their reduced dimensionality. In monolayer transition metal dichalcogenides (TMDCs), phase transformation is typically described as a martensitic process characterized by concerted atomic displacements. Nevertheless, the large energy barrier in 2D TMDCs makes such transformations difficult to realize, posing a substantial challenge to the experimental research on the microscopic mechanism, and hindering the precise regulation of material properties. To address this, we investigate the phase transformation in monolayer MoTe ₂ through advanced molecular dynamics simulations accelerated by deep learning potential. Our results uncover that the phase transformation proceeds in a one-dimensional (1D), domino-like manner, exhibiting features of both martensitic and reconstructive transformations. This unique mechanism provides tunability over the process, enabling remarkably enhanced nonlinear optical responses and rapid electrical switching. This work advances current phase transformation understanding and provides perspectives for the phase engineering in other 2D materials.