Phases and Architectures in Metal/Metal Oxide Systems Driven by Strong Metal–Support Interactions

Metal oxide supported metal catalysts are widely applied in industrial processes. Many of these materials dynamically evolve under reducing atmospheres, leading to metal nanoparticles partially or fully encapsulated by metal oxide shells, impacting catalytic performance. This phenomenon is known as strong metal–support interaction (SMSI) and is thermodynamically driven. However, understanding the metal/metal oxide interfaces derived from the broad and flexible compositional space and the large structural changes in SMSI structures is difficult to monitor experimentally. Here, we use density functional theory together with machine learning interatomic potentials and global minima optimization to investigate SMSI by building a set of interfaces between common catalytic metals (Ni, Pd, Pt) and reducible metal oxides (r-TiO2, CeO2, In2O3) at different reduction levels. Phase diversity arises from the competition between the formation of different metal oxides or binary alloys, while the local properties of the suboxide layers are responsible for the final architecture and composition determining the electronic properties of the material. Two descriptors related to the competition between alloy and oxide formation are proposed to elucidate the phase diversity. Our work provides a systematic approach to advance the design of SMSI-based catalytic materials by offering insights into the atomic-level architecture of the metal/metal oxide interfaces.