Intelligent Single-chip Si-FET based E-Nose using Metal and Oxides Functionalization

Noble metals and metal oxide semiconductors have been utilized for gas sensing applications with machine learning approaches and various hardware topologies. This paper presents a novel single-chip room temperature E-Nose integrating multiple gas sensing technologies to address cross-sensitivity, instability, and high-power consumption challenges. We deploy an array structure based on crystalline Si tandem FETs functionalized with metal oxide semiconductors (VO2, TiO2) and noble metals (Au, Pd) for enhanced sensitivity and selectivity. The fabricated E-Nose is electrically characterized and tested using a PDMS-based microfluidic channel in the presence of NH3, NO2, H2S and Air. The response of each FET to the gases varies due to the versatile sensing layers. This capability is further leveraged by developing a machine learning classifier based on Linear Discriminant Analysis. The trained model achieves 89.66% overall accuracy in gas classification without data pre-processing. This room-temperature, low-power, standalone, intelligent, multi-gas E-Nose showcases enhanced sensitivity and selectivity due to its unique hybrid nanomaterials, promising significant gas-sensing technology advancements.