近红外光谱法快速测定乌药中乌药醚内酯和去甲异波尔定含量

摘要

目的:建立乌药中乌药醚内酯和去甲异波尔定2种主要成分的近红外光谱定量模型.方法:采用高效液相色谱法测定乌药中乌药醚内酯和去甲异波尔定的含量,然后运用傅立叶变换近红外光谱技术并结合偏最小二乘法(PLS)、选择谱段为4 412 ～4 347,8720～4458cm-1,以“二阶导数+标准正则变换”对原始光谱进行预处理建立定量分析模型;采用校正模型的相关系数(r)、校正均方根误差(RMSEC)、内部交叉验证均方根误差(RMCECV)和预测模型相关系数(r)、预测均方根误差(RMSEP)对校正模型进行评价.结果:所建立的定量校正模型中乌药醚内酯和去甲异波尔定的r分别为0.996 14和0.985 55;RMSEC分别为0.004 36和0.036 1;RMCECV分别为0.040 11和0.171 74;预测模型乌药醚内酯和去甲异波尔定的r值分别0.917 0和0.903 4;RMSEP分别为0.010 5和0.090 1.结论:该研究建立的近红外定量模型稳定性较好、预测结果较准确,可用于乌药中乌药醚内酯和去甲异波尔定2种主要成分的同步快速测定.%Objective:To established a near-infrared spectroscopy quantitative model for simultaneous analysis of linderane and norisoboldine in Radix Linderae.Methods:First,the chemical values of linderane and norisoboldine in Radix Linderae were determined by high performance liquid chromatography.Fourier transformation near-infrared spectroscopy (FT-NIR) combined with partial least square (PLS) regression was used to establish and optimize the quantitative calibration model,choosing the spectrum ranges of 4 412 ～4 347 and 8 720 ～ 4 458 cm-1 and " Second derivative + SNV" mode.The calibration model was evaluated by the correlation coefficient (r),the root-mean-square error of calibration (RMSEC),and the root mean square of cross-validation (RMSECV) of the calibration model as well as the root mean square of prediction (RMSEP) of the prediction model.Results:For the quantitative calibration model the r,RMSEC,RMSECV of linderane and norisoboldine were 0.999 9,0.000 673,0.040 11 and 0.990 5,0.027 1,and 0.171 74,respectively.For the prediction model the r and RMSEP of linderane and norisoboldine were 0.917 0,0.009 54 and 0.903 4,0.076 3,respectively.Conclusion:The established near-infrared spectroscopy quantitative model is relatively stable,accurate and reliable for the simultaneous quantitative analysis of linderane and norisoboldine,and is expected to be used for the rapid determination of linderane and norisoboldine in crude drug of Radix Linderae.