(545b) Modeling of Nucleation, Growth, and Dissolution Kinetics of Paracetamol from Ethanol Solution for Unseeded Batch Cooling Crystallization with Temperature Cycling Strategy

Kim, Y. - Presenter, Georgia Institute of Technology
Rousseau, R., Georgia Institute of Technology
Kawajiri, Y., Nagoya University
Grover, M., Georgia Tech
Crystallization is widely used for industrial separations and purifications. Batch crystallization is often operated in a cooling mode, with the temperature gradually lowered to increase supersaturation and induce nucleation and growth. However, a temperature-cycling strategy, having both cooling and heating phases, can drive the system in new directions to achieve larger crystals or altered shape. Population balance models are often generated to describe crystal growth, but few models exist that describe the effects of both cooling and heating. Even less common is the comprehensive modeling of nucleation, growth, dissolution, and disappearance of crystals. However, it is this disappearance of small crystals during heating that enables the generation of larger crystals during subsequent cooling phases.

In this presentation, a one-dimensional population balance model is presented for unseeded cooling crystallization, with temperature-cycling strategies to control the mean crystal size and the crystal size distribution. This model can predict the path of crystallizing paracetamol from ethanol solutions based on the characteristics of primary nucleation, secondary nucleation, growth, dissolution, and disappearance of crystals. To collect solute concentration data, attenuated total reflectance-Fourier transform infrared spectroscopy (ATR-FTIR) was used, and to collect the crystal size distribution data, sieve analysis was carried out. The model was developed from two experimental and parameter-estimation steps, for dissolution and crystallization, to provide independent estimates of the model parameters. To solve the partial differential equation, we applied the conservation element/solution element scheme for the discretization of the spatial domain. The population balance model employed boundary conditions to describe the nucleation and the disappearance of crystals.

Dissolution experiments were carried out in isothermal conditions with different temperatures and different sizes of seed crystals. Unseeded crystallization experiments were performed beginning with a clear solution, with temperature cycling to control crystal size. The dissolution and crystallization models achieved good agreement with the experimentally measured volume density distribution of crystals as well as the concentration and supersaturation trends. These results demonstrate that the model can be used to inform future open- and closed-loop control strategies.