Abstract:Many heterogeneous chemical reactions involve gases catalyzed over solid surfaces at elevated temperatures and play a critical role in the production of energy, healthcare, pollution control, industrial products, and food. These catalytic reactions take place at the atomic level, with active structures forming under reaction conditions. A fundamental understanding of catalysis at the single atom resolution is therefore a major advance in a rational framework upon which future catalytic processes can be built. Visualization and analysis of gas-catalyst chemical reactions at the atomic level under controlled reaction conditions are key to understanding the catalyst structural evolution and atomic scale reaction mechanisms crucial to the performance and the development of improved catalysts and chemical processes. Increasingly, dynamic single atoms and atom clusters are believed to lead to enhanced catalyst performance, but despite considerable efforts, reaction mechanisms at the single atom level under reaction conditions of gas and temperature are not well understood. The development of the atomic lattice resolution environmental transmission electron microscope (ETEM) by the authors is widely used to visualize gas–solid catalyst reactions at this atomic level. It has recently been advanced to the environmental scanning TEM (ESTEM) with single atom resolution and full analytical capabilities. The ESTEM employs high-angle annular dark-field imaging where intensity is approximately proportional to the square of the atomic number (Z). In this Account, we highlight the ESTEM development also introduced by the authors for real time in situ studies to reliably discern metal atoms on lighter supports in gas and high temperature environments, evolving oxide/metal interfaces, and atomic level reaction mechanisms in heterogeneous catalysts more generally and informatively, with utilizing the wider body of literature. The highlights include platinum/carbon systems of interest in fuel cells to meet energy demands and reduce environmental pollution, in reduction/oxidation (redox) mechanisms of copper and nickel nanoparticles extensively employed in catalysis, electronics, and sensors, and in the activation of supported cobalt catalysts in Fischer–Tropsch (FT) synthesis to produce fuels. By following the dynamic reduction process at operating temperature, we investigate Pt atom migrations from irregular nanoparticles in a carbon supported platinum catalyst and the resulting faceting. We outline the factors that govern the mechanism involved, with the discovery of single atom interactions which indicate that a primary role of the nanoparticles is to act as reservoirs of low coordination atoms and clusters. This has important implications in supported nanoparticle catalysis and nanoparticle science. In copper and nickel systems, we track the oxidation front at the atomic level as it proceeds across a nanoparticle, by directly monitoring Z-contrast changes with time and temperature. Regeneration of deactivated catalysts is key to prolong catalyst life. We discuss and review analyses of dynamic redox cycles for the redispersion of nickel nanoparticles with single atom resolution. In the FT process, pretreatment of practical cobalt/silica catalysts reveals higher low-coordination Co0 active sites for CO adsorption. Collectively, the ESTEM findings generate structural insights into catalyst dynamics important in the development of efficient catalysts and processes.

Physics-Based Synthetic Data Model for Automated Segmentation in Catalysis Microscopy

An Atomistic Machine Learning Package for Surface Science and Catalysis

Chemical kinetics for operando electron microscopy of catalysts: 3D modeling of gas and temperature distributions during catalytic reactions

Simultaneous 2D Projection and 3D Topographic Imaging of Gas-Dependent Dynamics of Catalytic Nanoparticles

Visualizing single atom dynamics in heterogeneous catalysis using analytical in situ environmental scanning transmission electron microscopy

A deep learning approach for semantic segmentation of unbalanced data in electron tomography of catalytic materials

Data-efficient modeling of catalytic reactions via enhanced sampling and on-the-fly learning of machine learning potentials

Utilising Machine Learning for the Automated Characterisation of In Situ Electron Microscopy Experiments with Catalytic Systems

Synthetic Data Generation for Automatic Segmentation of X-ray Computed Tomography Reconstructions of Complex Microstructures

Visualizing Dynamic Single Atom Catalysis

Efficient Phase Segmentation of Light-Optical Microscopy Images of Highly Complex Microstructures Using a Correlative Approach in Combination with Deep Learning Techniques

Single Atom Dynamics in Chemical Reactions

Computational catalyst discovery: Active classification through myopic multiscale sampling

Learning heterogeneous reaction kinetics from X-ray videos pixel by pixel

Deep-learning-based prediction of nanoparticle phase transitions during in situ transmission electron microscopy

Semi-Supervised Segmentation of Mitochondria from Electron Microscopy Images Using Spatial Continuity

Machine learning of microscopic structure-dynamics relationships in complex molecular systems

Cascaded Diffusion Models for 2D and 3D Microscopy Image Synthesis to Enhance Cell Segmentation

Explainable Data-driven Modeling of Adsorption Energy in Heterogeneous Catalysis

Bridging microscopy with molecular dynamics and quantum simulations: An AtomAI based pipeline

A deep learned nanowire segmentation model using synthetic data augmentation