Fracture of 2D Materials – In situ Experiments and ML Parameterized Force Fields

2D materials are being employed in the development of next-generation electronics, optical, and sensor technologies, as well as in energy production and storage techniques, e.g., supercapacitors, solar cells, and battery electrodes. Such applications involve frequent mechanical deformations such as stretching and bending, so the lifespan (integrity and reliability) of the material is a critical feature. In this context, the abrupt and brittle failure of 2D materials require particular attention. In this presentation, I will discuss strategies for the in-situ electron microscopy fracture testing of 2D materials as well as advances in the parameterization of interatomic potentials (force fields) for accurate description of crack tips' atomic lattice reconstructions and bond dissociations. Experimentally, I will present two approaches: i) e-beam assisted crack propagation with measurement of crack tip deformation fields, based on atomic images, and ii) MEMS-based displacement-controlled fracture testing. The parameterization of force fields is based on a multi-objective genetic algorithm and machine-learning-inspired protocols, with training and screening data sets involving both equilibrium and far from equilibrium pathways such as phase transitions, vacancy formation energies, and bond dissociation energy landscapes. Using monolayer MoSe2 as a testbed, I will illustrate the effectiveness of the combined experimental-computational approach in measuring and predicting the toughness of the material, demonstrating in the process the advantages of ML-inspired force field parameterization in developing computational approaches with predictive capabilities.