Real-time monitoring of the column chromatography process of Ginkgo biloba using near-infrared and Raman spectroscopy combined with spectral fusion strategy

In the column chromatography elution process of traditional Chinese medicine (TCM), the real-time monitoring of active components plays an important role in product quality control. In this study, the column chromatography elution process in preparing of Ginkgo biloba leaf extract was performed in bench scale, where near-infrared (NIR) spectroscopy and Raman spectroscopy were collected to determine the concentration changes of several critical quality attributes. Based on single spectral data and fused spectral data, partial least squares and support vector machine models were established, respectively. The results indicated that for most analytes, models based on Raman spectra were superior to NIR spectral models, and fusion models established by the mid-level data fusion approach were superior to the single spectral models. The coefficients of determination (Rt2) of the optimal Raman and mid-level data fusion models were all higher than 0.96. The concentration profiles predicted by the mid-level data fusion quantitative models were in better agreement with the real trends defined by measured values. This is the first-ever study reporting the real-time monitoring of TCM column chromatography process by Raman spectroscopy combined with NIR spectroscopy, having vital reference values for process monitoring and understanding of the column chromatography for other TCM.