Chromatographic parameter estimation for the model-based design of a sugar separating simulated moving bed process

Rare sugars are becoming increasingly interesting for many industrial applications, due to properties such as low energy content and potential pharmaceutical uses. D-psicose is one of the rare sugars with a future use as a sweetener and precursor for pharmaceuticals. It can be enzymatically produced from a cheap substrate, Dfructose. Because of the position of the thermodynamic equilibrium the enzymatic reaction yields an epimeric mixture of substrate and product. To separate such sugar mixtures, chromatographic separation is the method of choice and for continuous separation simulated moving bed (SMB) technology has been established. In this work parameter estimation for the chromatographic separation of D-fructose/D-psicose using the recently established inverse method of chromatography investigating different chromatographic models. The isotherms were assumed to be linear for both sugar species. To account for limited the mass transfer in the system the linear driving force and the bilinear driving force models were investigated. The contributions of the axial dispersion were either lumped into a coefficient, λ, free for fitting or fixed by correlations derived by Chung & Wen and Gunn. The quality of the fittings was scored by calculation of the weighted residuals. The linear driving force model in combination with the lumped axial dispersion coefficient λ resulted in the best fit to experimental band profiles and were chosen for estimating the separation parameters at 20, 30 and 40 °C. The single column parameterization was used to design a SMB process. Purities in experiments performed at 20, 30 and 40 °C could with good accuracy be predicted by a true moving bed (TMB) model. Additionally, the model was used to perform a preliminary optimization.