摘要
目的:利用近红外漫反射光谱法测定黄芩、金银花、连翘、山羊角、熊胆粉5种中药提取物中的水分含量,并在此基础上总结出近红外漫反射光谱测定中药提取物中水分含量的一般规律。方法:近红外仪器参数设置为:分辨率4cm-1;扫描次数64次;光谱范围4000～10000cm-1;增益1×。分别选择包括水的3个特征波长(6900,5200,5180cm-1)的波段进行建模,并经多元散射校正,Savitzky-Golay平滑和一阶导数处理后,利用偏最小二乘回归方法建立5种提取物的近红外光谱与参考方法测得的水分含量之间的相关模型,并利用内部交叉验证和外部验证的方法对模型进行了优化。结果:所建的5个模型对验证集样品水分含量的预测值与实测值的相关系数均在0.90以上,预测误差均方根(RMSEP)均在0.05以下。利用所建模型对1批未知样品进行测试,结果满意。结论:该方法分析时间较短,预测结果准确,可推广应用于中药提取物含水量的快速测定。
Objective：Near infrared diffuse reflectance spectroscopy was used for measurement of moisture content in 5 kinds of traditional Chinese medicine extracts, namely Radix Scutellariae, Forsythia suspense, Flos Lonicerae, Cornu gorais, and Bear gall powder. And the general rule of determining the moisture content with N1R was summarized according to the experimental results. Method：The spectra were recorded with the average of 64 scans over the spectral region 4000 - 10000 cm^-1 at 8 cm^-1 intervals. The wavebands containing 3 characteristic wavelengths of water （6900,5200,5180 cm^-1） were selected, and were pretreated with Multiplicative Scatter Correction, Savitzky - Golay tilt, and first derivative. The calibration between the NIR spectra and the reference values of moisture content was obtained with Partial Linear Squares method and optimized through inner cross validation and external validation. Result：The coefficients of correlation of inner cross validation and external validation are both above 0. 90, and the RMSECV and RMSEP are both below 0. 05. The calibration models were used for testing a new set of samples, and the results were satisfied. Conclusion：The presented method is timesaving and accuracy, which indicates potential to be widely used in the moisture content determination of TCM extracts.
出处
《药物分析杂志》
CAS
CSCD
北大核心
2009年第10期1602-1606,共5页
Chinese Journal of Pharmaceutical Analysis
基金
浙江省重大科技专项社会发展项目(2008C13004-1)
新世纪优秀人才支持计划(No.NCET-06-0515)
关键词
近红外漫反射光谱
中药提取物
水分
near infrared diffuse reflectance spectroscopy
TCM extracts
moisture content
作者简介
第一作青Tel：（0571）88208428；E—mail：wshliwenlong@126．com
