拉曼光谱技术用于盐酸二甲双胍片含量的快速筛查

Rapid Determination of the Content of Metformin Hydrochloride Tablets by Raman Spectroscopy

目的:建立采用拉曼光谱技术测定盐酸二甲双胍片含量的方法。方法:以紫外法测定盐酸二甲双胍片的含量作为参考值,采集32批次样品的拉曼光谱图,结合偏最小二乘法建立定量分析模型。考察不同建模波段与不同光谱处理方法对所建模型预测精度的影响。结果:采用500.00～1 000.00cm^(-1)波段的光谱数据,结合多元散射校正与Savitzky-Golay平滑(7点3项式)处理后,所建定量模型预测精度高。模型相关系数(R)为0.990 1、校正均方差(RESEC)为0.160 1,"留一法"内部交叉验证相关系数(R)为0.942 6、交叉验证均方差(RESECV)为0.380 2,外部验证相关系数(R)为0.908 0、预测均方差(RESEP)为0.284 0,预测值与参考值的偏差均在±0.50%范围内。结论:该方法测定速度快、准确度高、简便环保,可用于盐酸二甲双胍片质量的现场快速筛查。

Objective: To develop a method for the content determination of metformin hydrochloride tablets by Raman spectroscopy. Methods: The content of metformin hydrochloride tablets was determined by UV as the reference value. Raman spectra were collected for 32 batches of samples. A quantitative model was established by combining Raman spectroscopy with partial least squares. The influences of different modeling bands and different spectral processing methods on the prediction accuracy of the model were investigated. Results: The optimized band for the modeling was 500.00-1 000.00 cm-1. Combined with multivariate scattering correction and Savitzky-Golay smoothing (7 point 3 term),the proposed quantitative model had high precision for the prediction. The model correlation coefficient R was 0.990 1,and RESEC was 0.160 1. The correlation coefficient of leave-one-out cross-validation R was 0.942 6,and RESECV was 0.380 2. External predictability of the model was adopted with the correlation coefficient R of 0.908 0,and RESEP was 0.284 0. The deviations between the predicted values and the reference values were within the range of plus or minus 0.50%. Conclusion: The method is fast,accurate,simple and environment friendly for the rapid content determination of metformin hydrochloride tablets.