近红外光谱法在白芍提取物纯化过程中快速质量控制研究

目的 建立一种基于近红外光纤透射光谱的白芍提取物纯化过程中的快速分析方法。方法 通过近红外透射光纤探头测定流经D-101大孔树脂的纯水洗脱液和70%乙醇洗脱液的近红外光谱（NIRS），并采用模式识别方法对两个过程进行定性判别研究。同时针对醇洗过程，以HPLC分析值作参比，采用偏最小二乘法建立醇洗过程中芍药苷和芍药内酯苷的定量校正模型。结果 模式识别方法可以准确无误地对水洗除杂过程完全程度进行监测，对醇洗脱过程中溶剂体系变化进行精确指示，同时所建定量模型也成功地用于预测相同批次和不同批次白芍提取物大孔树脂纯化过程的洗脱曲线，芍药苷和芍药内酯苷的预测均方差（RMSEP）分别为0.124、0.172和0.120、0.133，NIRS与HPLC测得值之间的相关系数（r）均在0.992以上。结论 本方法实时、高效、快速，可对纯化过程进行定性定量全面监控，可用于中药纯化过程的快速分析。

Rapid quality control in purifying process of Paeoniae Alba Radix extract using near infrared spectroscopy

Objective To establish a fast detection method during the purifying process of the extracts from Paeoniae Alba Radix based on near infrared optical fiber transmission spectroscopy. Methods The near infrared spectroscopy (NIRS) of the eluates by water and 70% ethanol which flowed through the macroporous resin D-101 was measured by NIR transmission optical fiber probe, and the method of pattern recognition was used to study the two processes. Aiming at the process of 70% ethanol elution, the contents of constituents were determined by HPLC as the reference values, and partial least square regression (PLSR) was applied to build the calibration models of paeoniflorm and albiflorin. Results The processes of water elution and 70% ethanol elution could be accurately monitored with pattern recognition, and the changes of the solvent system could be accurately indicated. The quantitative models were applied to successfully predict the elution curves of Paeoniae Alba Radix extracts from the same batch and different batches on macroporous resin. The root mean square errors of prediction (RMSEP) calculated by PLSR models for paeoniflorm and albiflorin were 0.124, 0.172, and 0.120, 0.133, respectively, and the correlation coefficient between the values calculated by NIRS and HPLC was all greater than 0.992. Conclusion The method is real time, efficient, and fast, and could be applied to the fast analysis for the purifying process of Chinese materia medica.