<![CDATA[Perfume is a cosmetic product that widely used by people to improve their appearance in social interactions. Perfume released specific fragrance from the essential oil. Manufacturers often mix the pure essential oils with hexylene glycol to reduce prices. Utilization of hexylene glycol as the solvent and diluent often reduce the odour profile of the perfumes. This paper investigated the development of an electronic nose (e-nose) based on a metal oxide semiconductor (MOS) gas sensor to detect hexylene glycol in perfumes. E-nose in this study was developed using MOS gas sensors from Figaro and Raspberry series, including TGS 822, TGS 826, TGS 2600, TGS 2620, MQ2, MQ3, MQ8, and MQ135. For the experiment, we collected 10 brands of commercial perfumes from the supermarket around Purwokerto, Central Java. All samples of perfumes were analysed using gas chromatography-mass spectroscopy (GC-MS) to detect the concentration of hexylene glycol in the samples. The concentration of hexylene glycol in the samples identified none (0%), low (1-20%), moderate (21%-50%) and high (more than 50%). Afterward, 10 brands of perfumes were separated into 15 samples, totally created 150 samples. All perfume samples were measured using an e-nose to obtain the responses. Analysis of sensor responses using principal component analysis (PCA) showed that e-nose was highly performed to discriminate the samples based on hexylene glycol concentration. Classification of 150 perfume samples using backpropagation neural networks (BPNN) grouped 150 perfumes in four different classes in which the accuracy of classification reached 96.36% for the training dataset and 92.50% for the testing dataset, respectively.]]>
Copyrights © 2024
