ABSTRACT
Near infrared spectroscopy(NIRS) was used for rapid quantitative analysis of saponins in Pien Tze Huang troches and powders. The near infrared spectra of Pien Tze Huang were collected,and the contents of notoginsenoside R1,ginsenoside Rg1 and ginsenoside Rb1 in Pien Tze Huang were determined by high performance liquid chromatography(HPLC) as the reference values. Then the near infrared spectra of the samples were associated with the reference values to establish the quantitative analysis models by using partial least squares(PLS) method. Finally,the models were verified by unknown samples. The results showed that root mean square error of cross-validation(RMSECV) of R1,Rg1,Rb1 and the total content was 0.095 1,0.555,0.414,0.960 mg·g-1 for the troches models,0.085 6,0.443,0.405,0.913 mg·g-1 for the powders models. After external validations,root mean square error of prediction(RMSEP) of R1,Rg1,Rb1 and the total content was 0.111,0.274,0.276,0.807 mg·g-1 for the troches models,0. 059 2,0. 322,0. 327,0. 705 mg·g-1 for the powders models. The averages of relative standard deviation between the predicted values and the chemical measured values were all less than 2.0%. According to the results of paired-t tests at the level of α = 0.05,there were no significant differences between the predicted values and the measured values. The established quantitative analysis models can be used to predict the contents of saponins in Pien Tze Huang accurately and the proposed method is simple,fast,non-destructive and environmentally friendly for the rapid detection and quality control of saponins in Pien Tze Huang.