On-Chip Integration of Self-Joule-Heated Oxide Nanosensors Discriminates Odor Components via Efficient Data...

Current data-based informatics approaches require efficient acquisition and processing of large data. Additionally, the collection and analysis of chemical information, involving combinations of molecules, require highly integrated sensors working in parallel. However, a trade-off exists between the collection of large data for high accuracy and minimizing the energy required to operate large arrays of sensors to detect the broadest range of analytes possible. The difficulty in achieving the high-density integration of heterogeneous sensing materials, which is inevitable in conventional chemical sensor arrays, is a critical bottleneck for efficient chemical data acquisition. Here we demonstrate the on-chip integration of self-Joule-heated oxide nanochannel sensors for discriminating complex odor components. Local temperature control of metal oxide surfaces via self-Joule-heating in nanofilm channels enables the individual modulation of chemical reactivity within each sensor of the integrated homogeneous-material sensor array chip. Herein, micrometer-scale temperature control and the modulation of molecular sensing properties were successfully implemented by optimizing the electrothermal design of SnO₂ nanofilm channel sensors. The fabricated sensor arrays can identify and classify complex odors from bananas and essential oils. Regression estimation of the ripeness (sugar-to-acid ratio) of bananas and the classification of four essential oils were demonstrated via simultaneous multitemperature operations of the homogeneous SnO₂ nanofilm channel sensor array. Feature analysis of the regression and classification models revealed the significant contribution of the sensor recovery time to the accuracy of data discrimination, highlighting that the desorption process modulation of molecules on the metal oxide surface via Joule heating plays a key role in complex odor discrimination.