Recent advances in machine learning guided mechanical properties prediction and design of two-dimensional materials

Abstract

Two-dimensional (2D) materials have garnered significant attention due to their exceptional physical properties and potential for diverse applications. Recent advances in the machine learning (ML) based methodologies have opened new avenues for predicting and designing the mechanical behaviors of these materials. This review comprehensively examines ML-based methodologies that predict key mechanical properties including fracture strength, elastic modulus, and Poisson's ratio, by integrating intrinsic structural features, defect characteristics, external loading conditions, and environmental influences. The ML-based ways of crack propagation and fracture mechanisms are thoroughly explored. Furthermore, high-throughput screening, reverse design and model+data based intelligence design of 2D materials are highlighted for accelerating the discovery and engineering of 2D materials with tailored mechanical properties. Emerging challenges such as the development of universal descriptors and potentials, multi-field coupling analysis, the integration of experimental, theoretical, and simulation datasets, and interpretable ML-based approach are discussed alongside future perspectives aimed at the design of high-performance 2D materials.