MSE Seminar: “Exploiting Automatic Image Processing and In-Situ Transmission Electron Microscopy to Understand the Stability of Supported Nanoparticles”

Date(s) - 02/07/2023
3:00 pm - 4:00 pm

Rhines Hall 125



Professor, Dept. of Materials Science and Engineering
University of Pennsylvania

Dr. Eric Stach is the Robert D. Bent Professor of Engineering in the Department of Materials Science and Engineering at the University of Pennsylvania and Director of the Laboratory for Research on the Structure of Matter, a National Science Foundation-sponsored Materials Research Science and Engineering Center. He received his B.S.E from Duke University, M.S.M.S.E. from the University of Washington, his Ph.D. in Materials Science and Engineering from the University of Virginia, and an M.B.A at Stony Brook University. He has held positions as Staff Scientist and Principal Investigator at the National Center for Electron Microscopy at the Lawrence Berkeley National Laboratory, then as Associate, and subsequently appointed Full Professor at Purdue University. Before his appointment at Penn, he was a Group Leader at the Center for Functional Nanomaterials, Brookhaven National Laboratory. He is a Co-founder and Chief Technology Officer of Hummingbird Scientific and is a Fellow of the American Physical Society, Materials Research Society, and the Microscopy Society of America.


The activity and lifetime of heterogeneous catalysts are linked with their structural stability in reactive environments. We have developed machine learning methods to track the temporal evolution of Au nanoparticles deposited on SiN as a model system to understand this process. We describe how systematic investigation of dataset preparation, neural network architecture, and accuracy evaluation lead to a tool for determining the size and shape of nanoparticles in high-pixel resolution TEM images. We use this algorithm to track nanoparticle coarsening, ripening, and sublimation as a function of time at elevated temperatures. We have developed an analytical model that describes this process, showing how local and long-range particle interactions through diffusive transport affect sublimation. The extensive data allows us to determine physically reasonable values for the model parameters, quantify the particle size at which Gibbs-Thompson pressure accelerates the sublimation process, and explore how individual particle interactions deviate from mean-field behavior. We observe that sublimation proceeds by sequential facet/defacetting transitions. We then utilize Kinetic Monte Carlo and Density Functional Theory to show how mobile adatoms form through desorption from low-coordination facets and subsequently sublimate. These results help to rationalize why evaporation rates vary between particles in a system of nearly identical nanoparticles.