Geographical discrimination of palm oils (Elaeis guineensis) using quality characteristics and UV-visible spectroscopy

Abstract

This preliminary study demonstrated the possibility of discriminating geographical origin of palm oils using conventional quality characteristics and UV-visible spectroscopy. A total of 60 samples, 20 from each region (North (N), South (S), and Central (C)) of Ondo State Nigeria, were analyzed for their quality characteristics and UV-visible spectra. Principal component analysis (PCA) and orthogonal projection to latent structure discriminant analysis (OPLS-DA) were applied to elaborate the data. Models were built on the most informative portion of the spectra (250–550 nm) as: untreated (without pretreatment) and standard normal variate—second-derivative-treated (SNV+2der) data matrices. OPLS-DA classification models were validated by independent prediction sets and cross-validation. PCA score plots of both chemical and spectral data matrices revealed geographical distinction between the palm oil samples. Significantly high carotene content, free fatty acids, acid value, and peroxide value distinguished Central palm oils. K extinction values, color density, and chlorophyll content were the most important quality parameters separating North oil samples. In the discriminant models, over 95% and 85% percent correct classification were recorded for spectral and chemical data, respectively. These results cannot be considered exhaustive because of the limited sample size used. However, the study suggested a potential analytical technique suitable for geographical origin authentication of palm oils with additional advantages that include the following: speed, low cost, and minimal waste.

This study presents a preliminary footprint for rapid geographical discrimination of palm oils using conventional quality characteristics and UV-visible spectroscopy. Both spectral and chemical data gave satisfactory predictions of regional differences among palm oils samples

http://ift.tt/2p13uVG