RESUMO
Avocado oil (AO) has been found to be adulterated by low-price oil in the market, calling for an efficient method to detect the authenticity of AO. In this work, a rapid and nondestructive method was developed to detect adulterated AO based on low-field nuclear magnetic resonance (LF-NMR, 43 MHz) detection and chemometrics analysis. PCA analysis revealed that the relaxation components area (S23) and relative contribution (P22 and P23) were crucial LF-NMR parameters to distinguish AO from AO adulterated by soybean oil (SO), corn oil (CO) or rapeseed oil (RO). A Soft Independent Modelling of Class Analogy (SIMCA) model was established to identify the types of adulterated oils with a high calibration (0.98) and validation accuracy (0.93). Compared with partial least squares regression (PLSR) models, the support vector regression (SVR) model showed better prediction performance to calculate the adulteration levels when AO was adulterated by SO, CO and RO, with high square correlation coefficient of calibration (R2C > 0.98) and low root mean square error of calibration (RMSEC < 0.04) as well as root mean square error of prediction (RMSEP < 0.09) values. Compared with SO- and CO-adulterated AO, RO-adulterated AO was more difficult to detect due to the greatest similarity in fatty acids' composition being between AO and RO, which is characterized by the high level of monounsaturated fatty acids and viscosity. This study could provide an effective method for detecting the authenticity of AO.
RESUMO
Raman spectra were used to distinguish waste cooking oil from edible vegetable oils. Signals at 869, 969, 1302 and 1080 cm-1 were found to be crucial to distinguish waste cooking oil from five edible oils using PCA. When waste cooking oil was added to soybean or olive oil, PCA could separate adulterated and pure oils, when the adulteration proportions reached 10% and 20%, respectively. Peaks at 969 (R2 > 0.951), 1267 (R2 = 0.987) and 1302 (R2 > 0.984) cm-1 responded linearly to adulteration. Heating assays and 1H NMR analysis revealed that differences between the Raman spectra of waste cooking oil and edible oils at 969 and 1267 cm-1 were directly related to heat treatment. This work highlights the potential for Raman spectroscopy to detect waste cooking oil.
