Experimental and theoretical analysis of solute redistribution during a progressive freeze concentration process

The performance of a progressive freeze concentration process depends on solute redistribution at the ice-liquid interface during the process, which, in turn, is characterized by the parameter ‘intrinsic partition coefficient’. A coupled heat and mass transfer model is proposed in this work to correlate this parameter to various characteristic velocities that are often encountered in a freeze concentration process. The robustness of the proposed model in predicting the final ice yield and the separation efficiency was validated through experimental trials conducted in a cylindrical stirred tank. Experiments investigated a model liquid solution (sucrose-water) with initial solute concentration ranging from 4% to 30%, stirring speeds varying between 100 and 500 rpm, and different cooling temperature profiles. Within the investigated characteristic velocity range (0.017–0.2), the correlation between characteristic velocity and intrinsic partition coefficient could be well approximated using a Sigmoidal function. A variation of 85% was achieved in the values of the intrinsic partition coefficient, confirming the limitations of a constant intrinsic partition coefficient, a common practice in the existing models. In addition, the proposed approach demonstrated an improvement in the prediction accuracy of the overall separation efficiency of the progressive freeze concentration process by about 40%.