Application And Modeling Studies On Chromatographic Processes In Separation And Purification Of Natural Substances

Chromatography has become one of the most widely used techniques in separation and purification of natural substances. However, the relevant studies on this technique are not comprehensive, especially on the fundamental studies. The goal of this paper is to study the chromatographic processes widely used in natural substances separation and purification, including both application and fundamental aspects. In application research, an economical method combining adsorption chromatography and crystallization was developed. In modeling research, the artificial neural network (ANN) was used to simulate the chromatographic processes. In this paper, the ANN-flxed bed model, ANN-reverse phase liquid chromatography model (ANN-HPLC) and ANN-high speed counter current chromatography model (ANN-HSCCC) were developed.In this paper, the separation of solanesol using fixed bed chromatography with macroporous resins was studied and then the solanesol was further purified by crystallization. By using this method, the purity of solanesol increased from 50% to 94.51%. The results showed that high-polarity solvent and low temperature were advantageous to the adsorption process; the stepwise elution was adopted in desorption process; the thermodynamic parameters such as enthalpy, Gibbs free energy and entropy changes were calculated and these values showed that solanesol adsorption process was exothermic and spontaneous. The results obtained in this part may provide scientific references for the large-scale solanesol production from tobacco leaves extracts.The general rate model including particle size distribution (PSD) and variation of isotherm (VOI) was developed. To examine the validity of the model, the theoretical predictions were compared with the experimental data obtained at different conditions. The results showed that the theoretical predictions were well consistent with the experimental data. The calculated results also showed that apparent differences can be observed between experimental data and the simulated results when only the PSD or VOI was taken into account. The theoretical predictions by model without considering PSD showed that the breakthrough occurred earlier and approached the plateau concentration much faster. The theoretical predictions by model without considering VOI showed that the breakthrough curve became less steeper and the time required to reach breakthrough point and plateau concentration was delayed.When developing the ANN- fixed bed model, the model was developed to describe the breakthrough behavior in fixed bed with macroporous resins. The encouraging simulated results showed that the ANN model could describe present system better than the modified general rate model. By using the predictive ability of ANN model, the influence of each experimental parameter was investigated. Predicted results showed that with the increases of particle porosity and the ratio of bed height to inner column diameter (ROHD), the breakthrough time was delayed. On the contrary, an increase in feed concentration, flow rate, mean particle diameter and bed porosity decreased the breakthrough time. When developing the ANN-HPLC model, an optimization strategy combining ANN and chromatographic response function (CRF) for chromatographic separation in HPLC was proposed. The ANN was used to simultaneously predict the resolution and analysis time, which are the two most important aspects in chromatographic separation. Subsequently, a CRF consisting of resolution and analysis time was used to predict the optimal operation conditions for different specialized purposes. The expected chromatograms were obtained at the predicted conditions, which verified the applicability of present method. Based on the results of this study, sequential combination of ANN and CRF can provide a more general, flexible and efficient optimization method for chromatographic separation.When developing ANN-HSCCC model, the effects of separation parameters on retention of stationary phase (S_f), resolution and retention time were studied. The ANN was used to simultaneously predict the effects of operation conditions and physical properties of two-phase systems on S_f. It was found that more accurate predictions were achieved by means of the ANN. Subsequently, a chemometrics approach combining Box-Behnken response surface model and Derringer's desirability function was applied for simultaneous optimization of resolution and analysis time in HSCCC. The merging of the two parameters was accomplished using the Derringer's desirability function with subsequent optimization by a Box-Behnken response surface design. The developed model was checked by statistical analysis. By implementing the optimal conditions predicted by the validated model, enhanced resolution between two similar analytes was achieved in a reasonable time. The analyses and results obtained in this paper will benefit to improve the efficiency of CCC separation.