No abstract

Preparative separation and purification of rebaudioside A from steviol glycosides using mixed-mode macroporous adsorption resins (MARs) were systematically investigated. Mixed-mode MARs were prepared by a physical blending method. By evaluation of the adsorption/desorption ratio and adsorption/desorption capacity of mixed-mode MARs with different proportions toward RA and ST, the mixed-mode MAR 18 was chosen as the optimum strategy. On the basis of the static tests, it was found that the experimental data fitted best to the pseudosecond-order kinetics and Temkin-Pyzhev isotherm. Furthermore, the dynamic adsorption/desorption experiments were performed on the mini column packed with mixed-mode MAR 18. After one run treatment, the purity of rebaudioside A in purified product increased from 40.77 to 60.53%, with a yield rate of 38.73% (W/W), and that in residual product decreased from 40.77 to 36.17%, with a recovery yield of 57.61% (W/W). The total recovery yield reached 96.34% (W/W). The results showed that this method could be utilized in large-scale production of rebaudioside A from steviol glycosides in industry.

The adsorption feature of different kinds of polystyrene-based macroporous adsorption resins (MARs) was investigated systemically at constant temperature employing Rutin as the adsorbate. Different from traditional adsorption patterns, Langmuir and Freundlich adsorption, and the results showed interesting aspects: (1) With the increase of the volume of the initial solution, the adsorption capacity increased to the maximum, and then decreased gradually. (2) Experimental results clearly verified the opinion that the adsorption process of MARs could be divided into three stages-macropores, mesopores, and micropores-by the capillary effects occurring at the two intersections, and the adsorption feature for every stage could be described well by the fourth type of Brunauer model. (3) The model that the inductive effect transmitted to the first layer could not interpret our experimental results reasonably. Thus, the model that the inductive effect passed on to a higher layer was proposed by investigating regression of the experimental results and the conclusion that the inductive effect transmitted to the third layer was drawn.