近红外光谱法在乌鸡白凤丸半成品质量控制中的应用

Application of near infrared spectroscopy for quality control of in-process Wuji Baifeng Pill

目的:探讨近红外光谱法在乌鸡白凤丸半成品(乌鸡、鹿角胶、鳖甲、桑螵蛸、人参等)质量控制中的应用。方法:采用近红外漫反射光谱技术结合化学计量学方法对乌鸡白凤丸半成品进行定性和定量分析。采集了95份同仁乌鸡白凤丸半成品和19份阴性(缺味)半成品的近红外漫反射光谱。以75份合格半成品的近红外光谱为标准,建立相似度匹配模型,并将其用于计算其余样品的相似度。采用偏最小二乘回归法建立了半成品中芍药苷和水分含量的定量分析模型,对未知样品进行含量预测。结果:相似度匹配模型能体现同仁乌鸡白凤丸半成品的质量特征,并准确反映药味的缺失。芍药苷和水分定量校正模型的相关系数分别为0.965 7、0.992 1,验证集均方差分别为0.005 5%、0.45%。结论:本方法准确可靠,操作简便,快速无损,可推广用于工业现场的在线检测。

AIM： To study the application of near infrared spectroscopy for quality control of in-process Wuji Baifeng Pill（ Callus Domesticus, Colla Comus cervi, Carapax Trionycis, Ootheca Mantidis, Radix et Rhizoma Ginseng, etc. ）. METHODS： Qualitative and quantitative methods for analysis of in-process product was developed by diffuse reflectance near infrared spectroscopy combined with chemometrics. Spectra of 95 samples of Tongren Wuji Baifeng Pill in-process and 19 raw material negative in-process products were collected. Similarity match model was built on the basis of spectra of 75 batches of qualified in-process products. Using this model, the similarity match values of other samples were calculated. Multivariate calibration models for paeonin and water content in inprocess product were developed with partial least square algorithm. RESULTS： The results showed that the proposed similarity match model could reflect the quality characteristic of Tongren Wuji Baifeng Pills in-process and reveal the absence and presence of raw materials. The correlation coefficients of the calibration model were within 0. 965 7 and 0. 992 1 while the values of root mean square error of prediction were 0. 005 5% and 0.45% , respectively. CONCLUSION： The experiment shows that the established method is accurate, fast and non-destructive, and can be applied to on-line detection.