Rapid Detection of CO, H₂S, and NO₂ Mixtures Using an Integrated SnO₂-Based Sensor Array Combined with Machine...

Area of Science:

• Materials Science
• Chemical Sensing
• Sensor Technology

Background:

• Industrial gas sensing demands precise detection of complex mixtures, exceeding single-sensor capabilities due to cross-sensitivity.
• Existing methods struggle with the speed and accuracy required for real-time monitoring of multiple gases.

Purpose of the Study:

• To develop a low-cost, fast detection system for mixed gases (CO, H2S, NO2) using an integrated sensor array.
• To leverage machine learning with time- and frequency-domain analysis for enhanced gas identification.

Main Methods:

• Designed a metal oxide semiconductor sensor array with distinguishable responses.
• Applied rapidly switched heating signals to microheaters.
• Extracted time- and frequency-domain features for machine learning model training.

Main Results:

• Achieved successful classification and concentration prediction of CO, H2S, and NO2 mixtures.
• Demonstrated high accuracy (96.30%) and determination coefficient (R^2=0.97) using only the first 20 seconds of adsorption data.
• Significantly reduced the data set size required for rapid mixed gas detection.

Conclusions:

• The integrated sensor array with machine learning offers a feasible solution for low-cost, rapid detection of ternary gas mixtures.
• This method advances the application of metal oxide semiconductor sensor arrays in industrial gas sensing.
• The approach enables efficient gas analysis, reducing data requirements and improving detection speed.