关键词: 巴戟天, 混伪品, 红外光谱, 聚类分析, 主成分分析

Abstract: In this study, infrared spectroscopy was adopted to analyze the infrared spectra of Morinda officinalis, Morinda umbellata, Kadsura coccinea and Alpinia officinarum. Spectral analysis revealed that the differences in spectral peak shapes and positions among the four medicinal materials were mainly concentrated in the ranges of 3600 ~ 3000 cm^-1 and 1800 ~ 600 cm^-1. Taking into consideration the results from second-derivative spectral analysis, it was confirmed that the characteristic peaks corresponding to starch, saccharides, calcium oxalate and esters existed in all four medicinal materials and the superposition of their characteristic peaks led to the spectral discrepancies among them. The cluster analysis and principal component analysis were further applied for classification research. Cluster analysis demonstrated that 40 random samples of Morinda officinalis, its adulterants and compatible herb could be divided into 4 categories, while principal component analysis classified these samples into four distinct regions. The above analysis results indicate that obvious fingerprint spectral differences exist in the infrared spectra of Morinda officinalis, its adulterants and compatible herb Alpinia officinarum, which enables the rapid identification and differentiation of the above herbal materials.

Key words: Morinda officinalis, adulterant, infrared spectroscopy, cluster analysis, principal component analysis