About this Abstract |
| Meeting |
2026 TMS Annual Meeting & Exhibition
|
| Symposium
|
Characterization of Minerals, Metals and Materials 2026 - In-Situ Characterization Techniques
|
| Presentation Title |
4D-STEM and Deep Leaning Analysis of Highly Disordered Materials |
| Author(s) |
Jinwoo Hwang |
| On-Site Speaker (Planned) |
Jinwoo Hwang |
| Abstract Scope |
Recent advances in 4-dimensional scanning transmission electron microscopy (4D-STEM) have provided new ways to characterize the nanoscale structure of heterogeneous materials in great detail. Despite its usefulness, the vast amount of data produced by 4D-STEM can be challenging to analyze using conventional methods. We demonstrate a new implementation of machine learning (ML) analysis of 4D-STEM data to study structural heterogeneity in metallic glasses (MGs). ML identifies the structural symmetry embedded in each 4D-STEM nanodiffraction pattern, revealing details of medium-range ordering that are key to understanding important properties of MGs. The results reveal substantial differences in structural heterogeneity as the composition changes in Zr-based MGs, which we connect to the proposed concept of ideal glass structures and their mechanical properties. Correlating this insight to atomistic and mesoscale deformation simulations reveals the important structural length scale relevant to shear transformation zones, local plasticity carriers that have been difficult to understand for decades. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Characterization, Mechanical Properties, Modeling and Simulation |