Classification and Identification of Volatile Organic Solvents based on Functional Groups using Electronic Nose

Abstract

The Metal Oxide Semiconductor gas sensors based on SnO2 indicate cross sensitivity to many volatile organic compounds. Therefore, this study is focused on developing a methodology to distinguish organic solvents based on the functional groups present using an array of Metal Oxide Semiconductor gas sensors. Here, representative compounds for aliphatic, aromatic hydrocarbons, carbonyl groups, esters, alcohols and dichloromethane were used to evaluate gas sensors. Then data were analyzed using Principal Component Analysis and k-Nearest Neighbor methods. Finally, k-Nearest Neighbor best model was developed to predict the chemicals based on the sensor data. The overall results of this study sufficiently explain that the developed electronic nose system can distinguish the chemicals tested with Principal Component Analysis (96.6 percentage) and can predict with kNearest Neighbor (k=5) (90 percentage) the chemicals based on the sensor responses. These results demonstrate that the developed electronic nose can be used to classify and identify chemicals based in different functional groups.