Influence of Viscosity on Variously Scaled Batch Cooling Crystallization from Aqueous Erythritol, Glucose, Xylitol, and Xylose Solutions.

This study presents a comprehensive comparison of the batch cooling crystallization performance of aqueous solutions containing sugars and sugar alcohols, namely, erythritol, glucose, xylitol, and xylose. Erythritol and xylitol are commonly used alternative sweeteners to replace sucrose. They can be obtained by fermentation-based bioprocesses, where glucose and xylose are typical raw materials. These model compounds were selected based on their differing rheological nature: saturated erythritol solution has a viscosity lower than 3 mPa·s, whereas xylitol has the highest viscosity: greater than 90 mPa·s in the studied temperature range. Viscosities and densities of saturated solutions as well as apparent viscosities of crystal-mother liquor suspensions were measured. The purpose was to evaluate their crystallization behavior within a specific temperature range from 40 to 20 °C and batch time of 2 h, with the aim of understanding the influence of viscosity on the process more comprehensively. The comparison within the selected compound systems was carried out in terms of the physical properties of the mother liquor and the crystalline product. In addition to empirical laboratory-scale (0.1 and 1 L) studies, larger-scale simulations (1 and 100 m3 ) were performed with the experimental data obtained on average particle size, density, and viscosity for mother liquor and crystal-mother liquor suspensions. Mixing characteristics, such as the dissipation energy, mass transfer coefficient, energy of collisions, and micromixing time, were calculated with VisiMix software when using a single or dual impeller mixer. Furthermore, the scaling up of erythritol, xylitol, glucose, and xylose batch cooling crystallization from 40 to 20 °C based on the scaling-up rule of constant tip speed and energy of dissipation was done with VisiMix to obtain overall data on mixing conditions with crystallizers of 1 and 100 m3 in volume. Furthermore, ANSYS CFD software was used to determine the strain rates close to the impeller tip and velocity profiles on various crystallizer scales.