Monitoring And Modeling Of The Purification Process Of Gardenia Extract Macroporous Resin

Macroporous resin column chromatography is an important purification technology for the pretreatment of traditional Chinese medicine.The purification process is essential to ensure the quality of traditional Chinese medicine.Understanding the characteristics and regularity of purification process has an important significance and value to ensure and improve the quality of traditional Chinese medicine.This paper takes macroporous adsorption resin purificating gardenia extract as the research object,near infrared spectroscopy technology was combined with chemometrics methods to study the characteristic of purification process changes.In addition,in this paper,the process model was built based on the parameters of macroporous adsorption resin purification process,and the fusion modle was built based on the process parameters data and near infrared spectrum data.The main research content includes the following two parts:1.Near infrared spectrum technology monitoring macroporous resin purification processNear infrared(NIR)spectroscopy technology was used to monitor and detect the end point of the elution process during the macroporous resin purifying the gardenia extract.The quantitative model was built to detect the concentration of geniposide in the alcohol eluate.Savitzky-Golay(SG)smoothing,first order derivatives,multiplicative scatter correction(MSC),standard normal variate transformation(SNV),wavelet denoise spectral(WDS)was compared to pretreat spectra,and interval partial least squares(siPLS)method was used to screening the best wavebands.Partial least squares algorithm was employed to build the quantitative model of concentration of genipinoside.Experimental results show that the optimal spectral pretreatment methods is SNV combination siPLS,the best modeling wavebands were 5805?6102 cm-1,6102?7309 cm-1 and 7911?8208 cm-1,respectively.At last,the RMSEP and RPD of the established model were 0.464 mg·mL-1and 6.90.According to the multivariate statistical process control(MSPC)theory and method,a new method for model updating was successfully established.The method can make full use of the near infrared spectrum data of calibration and validation sets and the HPLC data of calibration set.After adopting the new methods on the original model,the prediction effect is significantly increased,and the performance of the new method is better than that of the other four model update methods.The NIR spectroscopy quantitative analysis model of small scale was used to analysis alcohol elution phase of pilot experiments,and the sample data of pilot process was used to update the model.The results show that the quantitative model of small scale can directly predict the pilot process.Using qualitative analysis and quantitative analysis to determine the end point of gardenoside elution.The results show that the qualitative analysis method cannot accurately judge the end point of gardenoside elution.Quantitative analysis method can accurately describe the concentration changes of gardenoside of the elution phase,and according to the results of the quantitative model,the end point of gardenoside elution can be accurately judged.2.Process modeling of macroporous adsorption resin purificationTo study the macroporous adsorption resin purification process of gardenia extract,partial least squares(PLS)algorithm was employed to build the process model.Eight purification process parameters was used as the input variables to build process model with the elution volume of geniposide,gardenoside transfer rate,the percentage of geniposide,the elution volume of genipin 1-gentiobioside,genipin 1-gentiobioside transfer rate and the percentage of genipin 1-gentiobioside,respectively.The results showed that performance of the process model which was built only by using the the process parameters can not meet the application requirements.During the macroporous adsorption resin purification process,samples were collected in the loading sample phase,water elution phase and alcohol elution phase.Meanwhile,the the near infrared spectrum of these samples were obtained.Near-infrared spectroscopy and process parameters were combined to establish the integrated process model.By comparing different methods of data fusion,the final result shows that the three stages of full wave spectrum modeling effect is the best.Data fusion of near infrared spectrum data and process parameters can improve the performance of the process model.In conclusion,this paper has made a systematic research on the macroporous adsorption resin purification process.The NIR spectroscopy quantitative analysis model of elution stage was build.A new model updating method was established based on the MSPC theory.The end point of elution stage was determined by the quantitabive model.Meanwhile,this paper has study the date fusion modeling method,and the integratd model was build based on the near infrared spectrum data and process parameter date.