兰索拉唑片近红外漫反射光谱定量快速分析模型的建立
Construction of quantitative determination models for lansoprazole tablets analysis using near-infrared reflectance spectroscopy
分类号:
出版年·卷·期(页码):2013,33 (12):0-0
DOI:
10.16155/j.0254-1793.2017.01.01
-----摘要:-------------------------------------------------------------------------------------------
目的:采用近红外漫反射光谱分析技术和化学计量学的方法,建立兰索拉唑片快速、无损的含量测定分析模型。方法:以不同企业生产的兰索拉唑片为分析对象,通过聚类分析方法确定校正集和验证集,光谱处理方法为一阶导数+直线差减法;谱段范围为9746.9~7498.2 cm-1,6101.9~5446.2 cm-1;回归方法为偏最小二乘法(PLS)。结果:142批样品经内部交叉验证建立预测模型,浓度范围为4.72%~17.8%,决定系数(R2)为90.83%,交叉验证均方差(RMSECV)为0.730;用37批样品进行外部验证,外部验证均方差(RMSEP)为0.281,平均相对偏差为2.2%;测定未参与建模的4个厂家10批样品的含量,相对偏差小于3.5%。结论:用本文所建立的分析模型测定含量,结果准确,样品处理简单,可用于药品的现场快速分析。
-----英文摘要:---------------------------------------------------------------------------------------
Objective: To develop a near-infrared(NIR) method for rapid content determination of lansoprazole tablets. Methods: According to the cluster analysis method,the NIR spectra of lansoprazole tablets from different manufacturers were divided into calibration and validation sets.The quantitative model was established by partial least squares (PLS) algorithm with 1st derivative and straight line subtraction as the preprocessing method and 9746.9-7498.2 cm-1,6101.9-5446.2 cm-1 as the spectral ranges. Results: 142 batches of Lansoprazole tablet samples were analyzed,and the concentration ranged from 4.72%-17.8%.The root mean square errors of cross validation (RMSECV) was 0.730 and the determination coefficient(R2) was 90.83%.The root mean square error of prediction (RMSEP) of 37-batch samples in test was 0.281.The mean relative difference of prediction for test set was 2.2%.All the relative deviations of the true value and NIR prediction of 10 samples from 4 manufactures were all below 3.5%. Conclusion: The established method is simple,rapid,and accurate,which can be applied to on-site rapid drug testing.
-----参考文献:---------------------------------------------------------------------------------------
欢迎阅读《药物分析杂志》!您是该文第 871位读者!