感冒灵提取过程近红外光谱在线检测技术研究

On-line monitoring of extraction of coldrine by near-infrared spectrum

中药材提取过程是感冒灵颗粒剂生产过程的首个环节,对药品质量影响显著。该文在线采集感冒灵中药材提取过程近红外光谱图,以高效液相色谱和差重法为参照方法,采用一阶导数法处理近红外光谱,运用偏最小二乘回归法(PLSR)分别建立提取液中蒙花苷、绿原酸和固含量的定量校正模型。采用相关系数(r)、交叉验证均方差(RMSECV)、校正集均方差(RMSEC)和验证集均方差(RMSEP)等指标优化建模参数,考察模型性能。3种质量控制指标的模型相关系数均达到0.95以上,蒙花苷和绿原酸、固含量的RMSEC和RMSEP分别为0.010 4和0.009 47,0.009 34和0.142,0.055 5和0.008 42,在线分析所建模型,预测值与实际测定值相关系数均大于0.97,其预测相对偏差(RSEP)分别为8.14%,8.17%,9.86%。研究结果表明,利用近红外光谱技术可以实现感冒灵中药提取过程多指标的在线检测和实时监控,该技术可用于生产过程中质量控制,缩小中间体批次差异性,保证药品质量稳定性,也为后续的产品质量回溯提供了实时生产数据。

Extraction of the four Chinese herbals is the beginning step of the production process of coldrine granules and influences on drug quality significantly. In this paper,the on-line near infrared spectrum was collected during the extraction process of coldrine and then pre-processed by the first derivative. Partial least square regression（ PLSR） model was developed for the quantity indicators of linarin,chlorogenic acid and solid content,according to results of both HPLC and weight-loss as reference methods. The correlation coefficient,root mean square error of cross-validation（ RMSECV）,root mean square error of calibration（ RMSEC） and root mean square error of prediction（ RMSEP） were used to optimize model parameters and confirm their performance. Correlation coefficients of three quality control indicator models reached more than 0. 95. Values of RMSEC of linarin,chloroenic acid and solid content were0. 010 4,0. 009 34 and 0. 055 5,respectively. And the values of RMSEP were 0. 009 47,0. 142 and 0. 008 42,respectively. The models,built on-line analyze data,revealed that the correlation coefficients of predicted values and measured values were greater than0. 97 and values of RSEP of linarin,chloroenic acid and solid content were 8. 14%,8. 17% and 9. 86%,respectively. The results showed that the NIR method could achieve the on-line detection and real-time monitoring of multi-indexes during the extraction process of coldrine. The technology could be used for drug quality control in the process of practical production,reducing the batch differences and ensuring pharmaceutical quality stability. In addition,it could provide real-time production data for subsequent product quality backtracking.