基于紫外与近红外光谱技术的青风藤渗漉和萃取过程在线监测和终点判断方法研究

Study on online monitoring and endpoint judging methods for the percolation and extraction process of Sinomenii caulis based on ultraviolet and near-infrared spectroscopy technologies

目的:利用紫外光谱(UVS)与近红外光谱(NIRS)对青风藤的渗漉和萃取过程进行在线监测和终点判断。方法:采集渗漉过程中渗漉液的UV光谱,使用偏最小二乘回归法(PLSR)建立与盐酸青藤碱(SH)含量的定量校正模型,采用因子数、估计均方根误差(RMSEE)、决定系数R^(2)及Q^(2)、交叉验证均方根误差(RMSECV)和预测均方根误差(RMSEP)评价模型的拟合能力及预测能力;萃取过程同样使用PLSR建立UVS与NIRS的定量校正模型,同样采用因子数、RMSEE、R^(2)、Q^(2)、RMSECV及RMSEP作为评价模型的拟合能力及对未知样品的预测能力指标。结果:无预处理与经过一阶导数预处理的渗漉过程UVS模型R^(2)、Q^(2)、RMSEE、RMSECV和RMSEP均较为接近,有较强的拟合能力和预测能力;通过单独建模后,萃取过程的NIRS和UVS模型的SH浓度预测性能较好,UVS单独建立的水相模型RMSECV较小(0.21),NIRS单独建立的氯仿相模型RMSECV较小(2.68)。结论:利用UVS与NIRS技术,建立青风藤渗漉过程与萃取过程的SH浓度模型拟合预测性能和稳健性良好,有望实现渗漉过程和萃取过程的可视化。

Objective:Using ultraviolet spectroscopy(UVS)and near-infrared spectroscopy(NIRS)for online monitoring and endpoint determination of the percolation and extraction processes of Sinomenii caulis.Methods:Collecting UV spectra of the filtrate during the percolation process,using partial least squares regression(PLSR)to establish a quantitative correction model for the content of sinomenine hydrochloride(SH),principal components(PCs),root mean square error of estimation(RMSEE),determination coefficients R^(2) and Q^(2),root mean square error of cross validation(RMSECV)and root mean square error of prediction(RMSEP)were used to evaluate the fitting and predictive ability of the model;PLSR was used to establish quantitative calibration model for UVS and NIRS in the extraction process,and factor number,RMSEE,R^(2),Q^(2),RMSECV,and RMSEP were also used as indicators to evaluate the fitting and predictive ability of the model.Results:The performance of the UVS model for the percolation process without preprocessing was similar to that of the model with 1st derivative preprocessing,and the predictive ability was strong.After separate modeling,the NIRS and UVS models of the extraction process showed better predictive performance for SH concentration.The water phase model established by UVS alone had a smaller RMSECV(0.21),while the chloroform phase model established by NIRS alone had a smaller RMSECV(2.68).Conclusion:The use of UVS and NIRS technology to establish a SH concentration model for the percolation and extraction processes of Sinomenii caulisi has good predictive performance and robustness,and is expected to achieve visualization of the percolation and extraction processes.