Publication: Atr-ftir spectroscopic analysis for forensic differentiation of cooking oils
Loading...
Date
2026-02
Authors
Naderajan, Shamini
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The accurate classification of cooking oils is important for quality control, and forensic investigations such as food fraud and adulteration. In this study, attenuated total reflectance–Fourier transform infrared (ATR-FTIR) spectroscopy combined with chemometric analysis namely PCA and LDA was used to discriminate various type of cooking oils based on their spectral profiles. The type cooking oil samples analysed in this study were (i) black seed oil, (ii) canola + sunflower oil, (iii) canola oil, (iv) castor oil, (v) coconut oil, (vi) corn oil, (vii) olive oil, (viii) perilla seed oil, (ix) palm oil, (x) rice bran oil, (xi) sesame oil and (xii) sunflower oil. The spectrum of ATR-FTIR in both fingerprint region (1800–600 cm⁻¹) and the C–H stretching region (3050–2800 cm⁻¹) was not reliable due to overlapping spectral features in both spectral regions which made it harder to differentiate different types of cooking oil. Therefore, PCA which is an unsupervised method was performed where clear clustering of the cooking oils can be seen in the scatterplot of PC2 versus PC3 especially for black seed oil, perilla oil coconut oil and palm oil in the second region. The chemometric analysis was followed by a supervised method namely, LDA where the principal component (PC) score obtained from PCA for its respective region. Overall classification accuracy for fingerprint region was only 50.8% where misclassification of samples was observed in canola, sunflower, and corn oils due to having similar chemical structure. In contrast, the second region has very high accuracy for classification of cooking with 90.0% accuracy. Sunflower oil, canola oil and the blend of sunflower and canola oil was later re-classified as vegetable oil to improve the accuracy of classification. This has greatly improved overall accuracy to 91.7%. However, sesame oil remained as the most misclassified sample due to having very similar chemical structures to other type of cooking oils. The findings in this study demonstrate that ATR-FTIR spectroscopy coupled with PCA-LDA is a rapid, non-destructive method to discriminate various fresh cooking oils
Description
Keywords
-