Fast and portable single-sensor electronic nose for accurate quality assessment of extra virgin olive oil using a temperature-modulated generic gas sensor.

Abstract

Preserving the quality of extra virgin olive oil (EVOO), regarded as the best quality of olive oil categories in compliance with the International Olive Council (IOC), is mainly impeded by blending it with inferior or alternative edible oils. A portable, accurate, and rapid electronic nose was developed to assess the purity of EVOO blended with small amounts of virgin olive oil (VOO) in five different relative proportions. A thermal shock-induced (TSI) SnO₂ gas sensor was employed, obviating the necessity for a sensor array and eliminating multi-dimensional drift. The discriminative response time was optimized to 3.2 s, with consistent repeatability over three-month experiments. After preprocessing, feature selection was performed, followed by a combination of principal component analysis (PCA) and linear discriminant analysis (LDA), which effectively segregated olive oil clusters in 3D feature space. A k-NN classifier utilizing Euclidean distance achieved an exceptional 98 % accuracy for k = 7 in identifying the five binary proportions of EVOO and VOO, proving the capability of the designed e-nose to detect the quantity of complex odor mixtures. The setup's accuracy was subsequently assessed using the 25 validation data, yielding a precision of 92 %. The successful classification rate of the designed e-nose was attributed to the lower oxidative stability of pure VOO compared to pure EVOO, which led to the faster formation and decomposition of volatile organic compounds (VOCs) of VOO on the sensing pellet of the TSI gas sensor. This innovative e-nose shows great promise for industrial applications in recognizing the mixtures of complex odors.