December 1 - 6, 2024
Boston, Massachusetts

Event Supporters

2024 MRS Fall Meeting & Exhibit
CH05.11.01

Machine Learning-Driven Automated Aberration Correction on Scanning Transmission Electron Microscopes

When and Where

Dec 4, 2024
2:00pm - 2:15pm
Sheraton, Third Floor, Fairfax B

Presenter(s)

Co-Author(s)

Zijie Wu1,Matthew Boebinger1,Rama Vasudevan1

Oak Ridge National Laboratory1

Abstract

Zijie Wu1,Matthew Boebinger1,Rama Vasudevan1

Oak Ridge National Laboratory1
This abstract has been authored by UT-Battelle, LLC, under contract DE-AC05-00OR22725 with the US Department of Energy (DOE). The US government retains and the publisher, by accepting the article for publication, acknowledges that the US government retains a nonexclusive, paid-up, irrevocable, worldwide license to publish or reproduce the published form of this manuscript, or allow others to do so, for US government purposes. DOE will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan (http://energy.gov/downloads/doe-public-access-plan).<br/>Scanning Transmission Electron Microscopy (STEM) has become an indispensable tool for material science, enabling high resolution imaging and analysis of condensed materials on atomic scale. However, achieving optimal image resolution is often hindered by aberrations in the electron optics, and aberration correction is an unavoidable prerequisite step in the beginning of almost every STEM experiment. While physics-based algorithms are generally available in modern STEM control software to auto-correct high-order aberrations, significant human input is still required to manually correct low-order aberration parameters such as defocus and astigmatism, presenting a tedious labor for STEM experts and a significant barrier for those new to STEM. In this talk, we discuss a novel approach leveraging machine learning (ML) techniques to automate aberration correction in STEM. We build baseline neural networks (NN) to predict aberration coefficients by learning from large datasets of simulated ronchigrams; we then combine the trained NN models with optimization techniques to automatically adjust the aberration coefficients on microscope for optimal resolution. By automating the challenging and repetitive process of aberration correction, our method has the potential to lower the technical barrier of STEM experiments and allow for more efficient material characterization and discovery.<br/>Microscopy research was performed at the Center for Nanophase Materials Sciences at Oak Ridge National Laboratory, which is a US Department of Energy (DOE), Office of Science User Facility.

Keywords

scanning transmission electron microscopy (STEM)

Symposium Organizers

Miaofang Chi, Oak Ridge National Laboratory
Ryo Ishikawa, The University of Tokyo
Robert Klie, University of Illinois at Chicago
Quentin Ramasse, SuperSTEM Laboratory

Symposium Support

Bronze
Protochips
Thermo Fisher Scientific, Inc.

Session Chairs

Robert Klie
Marta Rossell

In this Session