Speaker: Dr. James M. LeBeau
Associate Director
Analytical Instrumentation Facility
North Carolina State University

Abstract:
Displacement of atoms away from their idealized unit-cell positions contributes significantly to material properties and functionality, e.g. around solute atoms in an alloy or off-centering in ferroelectrics. Methods to accurately, precisely, and directly quantify these picometer scale distortions are, however, lacking. While scanning transmission electron microscopy (STEM) provides direct imaging of atomic structure, for example, accurate and precise measurements had been prevented by sample drift and scan distortion. To address this limitation, I will introduce revolving STEM (RevSTEM). The method achieves sub-0.1% accuracy and picometer level precision, which enables the direct analysis of local structure.

Multiple case studies will be presented to demonstrate the capabilities of this new technique to characterize bulk, thin film, and nano materials. I will show how picometer level precision enables the direct observation of static atomic displacements correlated with local chemistry in complex oxide solid solutions.  These results are combined with density functional theory to quantify the distortion origins. I will also show how RevSTEM images can be used to accurately determine crystallographic parameters in real-space, and to determine the structural origins of spontaneous polarization in ferroelectric HfO2 thin films.  Turning to surfaces, I will show how in situ STEM can be used to directly study surface reconstructions as a function of thermodynamic variables. Towards addressing this challenge, I will discuss our implementation of a deep convolutional neural network to autonomously quantify large electron diffraction datasets. The approach is orders of magnitude faster than traditional least-squares and opens new avenues for electron microscopy automation.