Unraveling the atomic-scale deformation of metallic glasses

Metallic glasses have a disordered structure like traditional window glass and are exceptionally strong materials, similar to hard steels. They are seen as an ongoing revolution in solid materials and advanced materials in general. However, similar to window glass, metallic glasses tend to break suddenly and completely under pressure, rather than deforming gradually. The limited ability of metallic glasses to deform without breaking makes them less suitable for use as new structural materials. When stressed, they can develop weak spots called shear bands, which can lead to catastrophic failure. At the moment we do not fully understand how the structure of metallic glasses relates to the formation of shear bands. The exact mechanisms and sequence of events that lead to the formation of the shear bands are difficult to resolve experimentally because of the disordered structure of metallic glasses and the fast propagation speed and small dimensions of the shear bands. Understanding how these shear bands form and how they relate to the structure of metallic glasses is crucial to making stronger and more durable materials. This project aims to investigate these problems by combining computer simulations with advanced experiments. Using powerful microscopes and high-performance computers, we will look closely at how metallic glasses behave at the atomic level. We will compare what we see in the experiments with what the simulations predict to understand how these materials deform. The first goal is to develop new methods for predicting and understanding how metallic glasses deform, from the smallest atomic movements to larger-scale deformations. To this end, we plan to use our newly proposed atomic-level mechanism underlying the formation of shear bands, called the STZ-vortex model. We will then further develop, optimize and apply the STZ-vortex mechanism and follow in great detail the correlation between structure/property fluctuations and local atomic rearrangements in metallic glasses from small elastic excitations to the final fracture. One exciting aspect of this project is that, thanks to the high level of detail that the new microscopes can achieve, we will be able to compare the experimental results directly with atomistic simulations. The second goal is to predict the deformation of metallic glasses using artificial intelligence and, in particular, machine learning. By using powerful machine learning techniques together with various environmental descriptors, important advances can be made in establishing atomic-level structure-property relationships in metallic glasses. A synergistic combination of multi-scale simulations, machine learning techniques and advanced experiments will provide valuable insights into the design of metallic glasses with precisely tailored properties, leading to the development of new, tougher materials.