Single-Device Artificial Olfaction for H₂/CO Discrimination via Transient Multistep Thermal Modulation on a SnO₂-MEMS Sensor

The development of intelligent single-device artificial olfaction systems is essential for real-time monitoring of hazardous gases, especially in scenarios requiring selective recognition of chemically similar analytes such as hydrogen (H2) and carbon monoxide (CO). Herein, we propose a SnO2-based micro-electromechanical system (MEMS) gas sensor driven by transient multistep thermal modulation (MTM), enabling robust discrimination of H2 and CO without the need for sensor arrays. By programming the MEMS-integrated microheater with multiple temperature steps, multidimensional dynamic response signals were extracted and analyzed using machine learning algorithms, including logistic regression (LR), support vector machine (SVM), and K-nearest neighbors (KNNs). The tailored multistep heating modes modulated the surface redox kinetics and electronic transport behavior of the SnO2 sensing layer, enhancing gas-specific response features across different heating phases. Compared to isothermal operation, multistep heating not only improved the sensing performance but also generated richer signal patterns, achieving up to 98.4% classification accuracy using KNN. This work establishes a compact, low-power strategy for intelligent gas discrimination, providing new insight into the design of artificial olfaction based on a single thermally modulated MEMS sensor.