| Summary: | 碩士 === 國立屏東科技大學 === 食品科學系 === 92 === This study explored the rapid quality analysis of edible oils by Fourier transform infrared spectroscopy - attenuated total reflection (FTIR-ATR), complemented by multivariate analysis. The first part of the investigation was on the authentication of olive oil adulterated with soybean oil, and the evaluation of constructed calibration model. Principal component analysis (PCA) was carried out to eliminate the multi-collinearity among multiple peak absorptions. 4 principal components were selected which retained 91.39% of the original variance. Calibration model was constructed with PCA followed by subsequent multiple linear regression (MLR). The determination coefficients (R2) for calibration set and validation set were 0.991 and 0.998, respectively. The standard errors for calibration set (SEC) and validation set (SEP) were 2.87 and 1.003, respectively. From scatter plot of residuals and other model evaluation parameters, it showed FTIR-ATR could effectively predict the adulteration of olive oil by soybean oil.
The second part dealt with the quantification of trans fatty acid in edible oils. Using trans fatty acid free olive oil as the reference, the linear regression model was constructed based on the samples spiked with pre-determined percentages of methyl elaidate, and the corresponding absorption at wavelength of 10.3 μm (966 cm-1). From the FTIR results complemented by gas chromatographic examinations, R2 of calibration model was 0.9936, and showed consistent predictive ability even at low concentration (2.4%). Overall, FTIR-ATR exhibited the advantages of speed and convenience, also was qualified as superior analytical tool for edible oil quality.
|
|---|
