Thermally activated excess noise by subgap density-of-states in Si-doped ZnSnO thin-film transistor-type gas sensor.

Amorphous oxide semiconductor (AOS) thin-film transistor (TFT) platforms have shown exceptional promise for advanced gas-sensing applications due to their intrinsic sensitivity, tunable electrical properties, and suitability for integration with flexible substrates. However, their practical sensing performance-particularly stability and limit of detection (LoD)-is fundamentally constrained by low-frequency noise (LFN), which destabilizes sensor baseline signals and diminishes the signal-to-noise ratio (SNR). Here, we comprehensively investigate the temperature-dependent LFN characteristics of silicon-doped ZnSnO (SZTO) TFT sensors. By employing energy-resolved subgap density-of-states analysis, we reveal that deeper donor-like states become thermally activated at elevated temperatures, significantly amplifying excess 1/f noise, especially within low-bias operational regimes critical to sensing. Such excess noise, which cannot be captured by conventional noise models, critically reduces the effective SNR, thereby degrading the minimum detectable gas concentration. Our results establish a direct correlation between temperature-induced activation of subgap trap states and sensor-relevant LFN behavior, providing essential guidelines for optimizing operating temperature and bias conditions to minimize electronic noise and maximize detection stability in real-world sensing environments.