| About this Abstract |
| Meeting |
2024 TMS Annual Meeting & Exhibition
|
| Symposium
|
Novel Strategies for Rapid Acquisition and Processing of Large Datasets from Advanced Characterization Techniques
|
| Presentation Title |
Understanding Relaxation Dynamics Beyond Equilibrium Using AI-Informed X-ray Photon Correlation Spectroscopy |
| Author(s) |
James P. Horwath, Xiao-Min Lin, Hongrui He, Qingteng Zhang, Eric M Dufresne, Miaoqi Chu, Subramanian K.R.S. Sankaranarayanan, Wei Chen, Suresh Narayanan, Mathew J. Cherukara |
| On-Site Speaker (Planned) |
James P. Horwath |
| Abstract Scope |
X-ray photon correlation spectroscopy (XPCS) is a synchrotron characterization technique which is unique in its ability to capture dynamics in out-of-equilibrium systems on μs-hour time scales and nm-μm length scales using two-time correlation functions (C2). However, the complexity and variability of C2, coupled with the lack of theoretical understanding of non-equilibrium systems, makes establishing connections between dynamics and material properties difficult. Improvements in detection and computing hardware, along with the commissioning of next-generation light sources, compound data analysis problems by drastically increasing the rate of data production. To address the challenges of data reduction and physics extraction from large experimental datasets we have developed an unsupervised deep learning framework for automated classification and interpretation of relaxation dynamics in complex fluids using XPCS C2. We will demonstrate how this approach is used to consider the full distribution of material behavior in experimental data, guide physical model selection, and quantify complex dynamics. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Characterization, Machine Learning, Modeling and Simulation |