R. Cai, Shisheng Wang, Y. Meng
Jul 26, 2012
A fast, simple and accurate near infrared spectroscopy (NIRs) method was investigated for the geographical origin discrimination and main flavonoid content determination of Chinese propolis. 180 propolis samples from four provinces of China were collected for establishing partial least squares (PLS) models of rutin, myricetin, quercetin, kaempferol, apigenin, izapinin, pinocembrin, chrysin, galangin and total flavonoids. The root mean square error of prediction (RMSEP) and correlation coefficient (R2) in the test set of the optimum models were 0.479 to 1.332% and 76.23 to 96.68%. The residual prediction deviation (RPD) values were all larger than 2.93. A paired t-test result showed that there was no significant difference between NIRs and HPLC methods at a 95% confidence interval. A principal component analysis (PCA)–Mahalanobis distance (MD) method was applied to classify the geographical origin of propolis samples, and 100% accuracy was obtained. Finally, a quality assessment was performed, and the propolis samples from Zhejiang and Liaoning province were classified as good quality. The developed NIRs method was a suitable tool for the rapid analysis of Chinese propolis, and might be also extended to the on-line quality control of propolis raw material in pharmaceutical industries, or monitoring industrial natural products extraction processes.