近红外光谱分析技术用于硫酸羟氯喹原辅料混合均匀度在线定量监测

Near-infrared spectroscopy for online quantitative monitoring of blend uniformity of hydroxychloroquine sulfate raw and auxiliary materials

目的建立在线监测硫酸羟氯喹原辅料混合均匀度的定量分析模型,以准确、快速判断混合终点。方法制备硫酸羟氯喹标示百分含量为70%~130%的原辅料混合物。采集近红外光谱,对原始光谱进行标准正则变换、一阶导数和Norris平滑处理,建模波段为8 372~9 045 cm^1、5 616~6 058 cm^1,运用偏最小二乘回归建立定量分析模型。运用建立的定量分析模型预测硫酸羟氯喹在原辅料混合过程中的标示百分含量,以高效液相色谱法(HPLC)作为参考方法对混合终点进行验证。结果建立模型所用主因子数为5;模型的校正误差均方根为0.96,校正集相关系数Rc为0.998;预测误差均方根为0.97,验证集相关系数Rp为0.998;交互验证的校正误差均方根为1.56,交互验证相关系数Rcv为0.995。近红外模型的预测结果与HPLC验证结果相符。结论所建近红外定量分析模型可以用于硫酸羟氯喹原辅料混合均匀度的在线定量分析,能够准确、快速判断混合终点。

Objective To establish a quantitative analysis model for online monitoring of the blending uniformity of hydroxychloroquine sulfate raw and auxiliary materials, so as to accurately and quickly determine the blending endpoint. Methods A mixture of excipients and hydroxychloroquine sulfate raw material was prepared with a labeling percentage of 70%-130%. The near-infrared spectrum was collected;and the standard normal variate transformation and first derivative by Norris smoothing were used for spectra pretreating, with 8 372-9 045 cm^1, 5 616-6 058 cm^1 used as the spectral bands. A quantitative analysis model was developed using partial least squares regression. The quantitative analysis model was used to predict the labeling percentage of hydroxychloroquine sulfate in the blending process of raw and auxiliary materials, and the blending endpoint was verified by high-performance liquid chromatography(HPLC) analysis. Results Five primary factors were used to establish the model. The root mean square error of calibration was 0.96 and the correlation coefficient of calibration set(Rc) was 0.998. The root mean square error of prediction was 0.97 and the correlation coefficient of validation set(Rp) was 0.998. The root mean square error of cross-validation was 1.56 and the correlation coefficient of cross-validation(Rcv) was 0.995. The prediction results of the near-infrared model was consistent with the results of HPLC verification. Conclusion The near-infrared model in this study can be used for online quantitative analysis of the blending uniformity of hydroxychloroquine sulfate, and it can accurately and quickly determine the blending endpoint.