Selectivity Optimization of GDC-Based Mixed-Potential Volatile Organic Compound Gas Sensors

Area of Science:

• Materials Science
• Chemical Sensing
• Machine Learning

Background:

• Sensor selectivity is crucial for accurate gas detection.
• Compounding materials and integrating machine learning can improve sensor performance.
• Nickel ferrite (NiFe2O4) and zinc ferrite (ZnFe2O4) are promising semiconductor materials for gas sensing.

Purpose of the Study:

• To enhance the selectivity of gas sensors by creating composite materials and utilizing machine learning.
• To investigate the gas-sensing properties of NiFe2O4:ZnFe2O4 composites for various volatile organic compounds (VOCs).
• To evaluate the effectiveness of machine learning models in improving gas recognition accuracy.

Main Methods:

• NiFe2O4 and ZnFe2O4 were synthesized using the sol-gel method.
• Composite materials were prepared by mixing NiFe2O4 and ZnFe2O4 at different mass ratios.
• Gas sensing performance was evaluated against five different VOC gases.
• Machine learning models (including Random Forest) were trained using sensor response data.

Main Results:

• The NiFe2O4:ZnFe2O4 composite sensor showed significantly increased response to five VOC gases.
• The optimal NiFe2O4:ZnFe2O4 ratio of 1:2 demonstrated excellent sensing performance.
• A detection range for Triethylamine (TEA) from 0.01 to 100 ppm was achieved with the 1:2 ratio sensor.
• A 1:3 ratio sensor exhibited an exceptionally wide TEA detection range (2 ppb to 200 ppm).
• Machine learning, particularly Random Forest, achieved a 94% recognition rate for the five gases.

Conclusions:

• Compounding NiFe2O4 and ZnFe2O4 effectively enhances gas sensor selectivity.
• Machine learning integration provides a powerful tool for improving sensor accuracy and selectivity.
• The developed composite sensors show potential for practical applications in gas detection.