About this Abstract |
Meeting |
2023 TMS Annual Meeting & Exhibition
|
Symposium
|
Characterization of Materials through High Resolution Coherent Imaging
|
Presentation Title |
Method Developments for High-efficient X-ray Coherent Diffraction Imaging |
Author(s) |
Yudong Yao, Junjing Deng, Henry Chan, Jeffrey Klug, Yi Jiang, Barbara Frosik, Zhonghou Cai, Ross Harder, Barry Lai, Mathew Cherukara |
On-Site Speaker (Planned) |
Yudong Yao |
Abstract Scope |
X-ray coherent diffraction imaging (CDI) has gained tremendous success in providing nanoscale characterization in materials science, chemistry, solid-state physics, and biology communities. Combined with Bragg diffraction and/or scanning modalities (ptychography), one can achieve three-dimensional imaging of lattice strains and large field-of-view imaging of extended samples. As a scanning variant of CDI, ptychography imaging speed is currently limited by the available coherent flux and the scanning mechanism. Here we report on our recent developments in ptychography imaging technique and the improvement of reconstruction methods to increase the imaging throughput. In addition to the data acquisition speed, the data processing speed of CDI is determined by the phase retrieval algorithms, which are traditionally iterative and are therefore computationally expensive. Our recent development of an unsupervised physics-aware neural network (AutoPhaseNN) has shown great advantages in accelerating the data inversion in CDI, potentially enabling real-time imaging capabilities. |
Proceedings Inclusion? |
Planned: |
Keywords |
Machine Learning, |