Application Of Near Infrared Spectroscopy In Quality Evaluation Of Panax Notoginseng

Objective: In the pre-feeding stage of Chinese herbal medicine manufacturing process,there is no method to evaluate the quality of raw materials by on-site rapid detection.At the same time,the change of environment and temperature will directly affect the final product quality in the production process.It needs a process analysis technology that can realize on-line real-time monitoring of intermediate product quality.Near Infrared Spectroscopy(NIRS)is a fast and non-destructive analysis technology,which can be used for on-site and on-line rapid analysis.Based on the above background,this study used Sanqi as raw material,using near infrared analysis technology,combined with Pharmacopoeia standards,established a rapid detection method for Sanqi quality detection,and established an online monitoring and quality control method for Sanqi extraction process,to meet the innovative development needs of pharmaceutical enterprises' intelligent manufacturing.Methods: 1.NIRS and primary data detection: 53 batches of Sanqi root samples were collected,and the contents of moisture,ethanol extract and saponinss in Sanqi root were determined according to the Chinese Pharmacopoeia 2015.2.Establishment of near-infrared quantitative off-line model: Three near-infrared quantitative analysis models of Sanqi moisture,ethanol extract and saponins were established by combining the results of traditional methods with NIRS.3.Establishment of near-infrared qualitative off-line model: 120 batches of samples of Sanqi were collected,including main root,branch root,velvet root,rhizome,flower and leaf.The identification of all parts of Sanqi was studied by using high performance liquid chromatography(HPLC)and NIRS combined three chemometrics methods.4.Establishment of near-infrared on-line quantitative model: Six batches of Sanqi were collected to simulate industrial extraction and production in the laboratory.The content of Sanqi saponins in the extraction process was monitored by ultra-high performance liquid chromatography(UPLC)combined with NIRS in real time,and an on-line near-infrared monitoring model was established.Result: 1.53 batches of Sanqi main root moisture content ranged from 8.67% to10.81%,ethanol extract content ranged from 16.00% to 22.97% and saponins content ranged from 5.85% to 9.27%,which all met the Pharmacopoeia standards.2.The correlation coefficients of calibration sets of NIRS quantitative models for moisture,ethanol extract and saponins content in the main roots of Sanqi were0.9972,0.9624 and 0.9951;the root mean square errors of calibration were 0.0396,0.7760 and 0.0698;the correlation coefficients of verification set were 0.9624,0.9884 and 0.988,and the root mean square errors of prediction were 0.173,0.595 and 0.134.3.The content of saponins in different parts of Sanqi was determined by HPLC.The content range of saponins in main roots overlapped with the parts of branch root and velvet root.NIRS combined with principal component analysis(PCA)could not distinguish the main roots,branch root and velvet root;NIRS combined with partial leasts quares discrimination analysis(PLS-DA)could identify all parts of Sanqi;NIRS combined with cluster analysis could not identify the main roots and branch root,and there was a classification error in the velvet root.4.In the model of on-line monitoring Sanqi extraction process by NIRS combined with partial least squares(PLS),the correlation coefficient of correction set was 0.9857,the root mean square errors of calibration was 0.405,the correlation coefficient of verification set was 0.9874,and the root mean square errors of prediction was 0.433.Conclusion: 1.The NIRS quantitative model of moisture,ethanol extract and saponins content of Sanqi established in this paper can be used for rapid detection of corresponding indicators.2.NIRS combined with PLS-DA was the best method for identification of all parts of Sanqi,which could be used for rapid location classification of Sanqi.3.NIRS could reflect the change of saponins content in Sanqi extracting process.The relative deviation of predicted results met the requirement of on-line detection of active ingredients in traditional Chinese medicine.It was suggested that enterprises built near-infrared system in actual production process.