Preparation and Optimization of Mesoporous SnO2 Quantum Dot Thin Film Gas Sensors for H2S Detection Using XGBoost Parameter Importance Analysis

Abstract

Tin oxide (SnO2) is a traditional gas-sensitive semiconductor with excellent response to various gases. However, its sensor performances are attenuated by the utility factor during gas diffusion in the sensing body. Therefore, the rational design of microstructure of devices is attractive and necessary because it may provide a sensible and controllable microstructure, which facilitates gas diffusion and inhibits the utility factor. Herein, the mesoporous tin oxide (MPTD) quantum dot thin film for H2S gas sensors is prepared by a facile route, which creates a mesoporous microstructure for thin films by the thermal decomposition of NH4Cl. The pore size of the thin films is controlled to be 19.36–40.13 nm. The mesoporous microstructure exhibits enhanced gas-sensing properties amounting to a 30-fold increase in response and 1/3 reduction in recovery time in H2S detection at room temperature (25 °C), with a limit of detection of 0.4 ppm. To determine the importance of sensor parameters such as pore size, film thickness, and grain size, an eXtreme Gradient Boosting (XGBoost) algorithm model was developed to examine the feature importance of each parameter on the gas-sensing performance of the MPTD sensors. The visual illustration of parameter importance is revealed to facilitate the optimization of technical preparation parameters as well as the rational design of semiconductor gas sensors.