In this work, a new device structure mixed potential gas sensor constructed with two sensitive electrodes, one counter electrode and one reference electrode was fabricated for Volatile Organic Compounds (VOC) classification. The construction of two sensitive electrodes enables one sensor to produce two outputs signal for the analyte. And the gas-sensitive performance of the sensor could be adjusted by applying different bias voltage on the counter electrode. The above methods could generate more sensing features, which makes it possible to replace sensor array with a single sensor. The output signal was further feature extracted and used to train the K-Nearnest Neighbor (KNN) machine learning model. The results of the classification shows that constructing a double electrode and applying a bias voltage could significantly enhances the discrimination ability of the sensor for different gases. By designing dual electrodes and counter electrode, a single sensor is realized instead of sensor array. It is important for the realization of artificial olfaction.
Read full abstract