Hillert Materials Modeling Colloquium series XXVI: Surface Energetics and Growth Mechanisms in Oxide Nanostructures
Professor Denis Music examines how tiny oxide structures form and how they can power next-generation thermoelectric devices and sensors. His work combines theory, experiments, and advanced simulations. In this seminar, he highlights how nanostructured oxides grow, how their surfaces and interfaces behave, and how machine-learning methods can model these processes efficiently. He also discusses how surface forces and growth dynamics shape the complex nanostructures that enable future advanced materials.
Time: Tue 2026-01-20 15.00 - 16.00
Location: Zoom
Video link: https://kth-se.zoom.us/j/61877868884
Language: English
Participating: Denis Music
Nanostructures are not only captivating for fundamental research due to their exceptional physical and chemical properties, which clearly differ from those of bulk materials, but they also play a crucial role in various innovative applications. Here, Ru-O, Nb-O, and Sn-O nanostructured oxides are explored, focusing on their growth mechanisms, surface and interface characteristics, and potential applications. For Ru-O nanostructures, a density functional theory (DFT) model explaining the formation of RuO2 nanorods is presented, emphasizing the pivotal role of Ru hyperoxides as nucleation sites. Hyperoxide clusters readily adhere to RuO2 surfaces, and adatoms impinging on these islands facilitate nanorod growth due to Ehrlich-Schwoebel barriers exceeding 22 meV per atom, an energy threshold sufficient to drive nanostructuring. Nb-O nanostructures form self-assembled NbO2 nanoslices that compete with formation of nanorods. DFT-based molecular dynamics reveal that (110) surfaces favor nanoslices due to their lower surface energy, while (001) surfaces stabilize nanorods above room temperature. This competition between surface configurations explains the observed texture modulation in NbO2 nanostructures. For Sn-O nanostructures, nanodendritic morphologies are observed. Capturing their evolution at the DFT level is computationally demanding; therefore, artificial neural networks trained on small-cluster interactions with SnO surfaces were employed. This machine-learning approach enables significant computational acceleration while maintaining predictive accuracy.
Together, these studies provide a comprehensive understanding of oxide nanostructure formation across multiple systems, highlighting the interplay between surface energetics, growth dynamics, and computational modeling in determining their morphology and stability.
