The limited selectivity of metal oxide semiconductor (MOS) sensors poses a persistent challenge to their practical implementation. Here, we demonstrate switchable CO/H2 detection through tailored gas adsorption properties enabled by compositional engineering of SnO2-Co3O4 nanocomposites. To address gas mixture quantification, we developed a dual-functional sensor array featuring distinct SnO2-Co3O4 sensing units optimized for H2 and CO detection, respectively. This architecture, when integrated with gradient boosting regression (GBR) algorithms, achieves simultaneous concentration prediction of both analytes in complex gaseous environments. Furthermore, a comprehensive investigation combining experimental characterization and first-principles calculations reveals the underlying mechanism governing selectivity modulation in these heterostructures. The work offers critical experimental insights and theoretical foundations for rational design of high-selectivity gas-sensing platforms.
Keywords: dual-sensor system; first-principles calculations; gas sensors; machine learning algorithms; selectivity.