Facet-Programmed Oxide Nanointerfaces for Metabolite Activation and AI-Assisted Disease Screening

Metal oxides are important nanocatalytic materials for small-molecule activation, owing to their variable valence states and efficient charge-transfer capabilities. This activation is governed by coupled oxide-molecule interfacial processes, for which conventional descriptors such as morphology, composition, and defects remain insufficiently precise. Here, we establish crystal facets as descriptors for oxide surface regulation and use laser desorption/ionization mass spectrometry (LDI-MS) as a high-throughput readout of metabolite activation. Across CeO₂, Co₃O₄, and Cu₂O, exposed facets strongly regulate metabolite desorption/ionization, with the (110) facet consistently exhibiting superior activity. Mechanistic studies and density functional theory calculations reveal that facet exposure modulates surface valence states, oxygen-vacancy formation, charge transfer, carrier dynamics, and energy dissipation. In particular, (110) CeO₂ integrates favorable defect structures, low work function, efficient charge separation, and localized energy confinement, enabling enhanced metabolite activation with up to three orders of magnitude signal improvement, high reproducibility, and salt/protein tolerance. The optimized (110) CeO₂ matrix enables serum metabolic fingerprinting for AI-assisted discrimination of hepatocellular carcinoma from chronic liver disease and biomarker identification. This work demonstrates facet engineering as a strategy for designing oxide nanointerfaces that translate molecular activation into chemically informative fingerprints for disease screening.