David Widenski

A qualitative comparison between population balances and stochastic models for non-isothermal antisolvent crystallization processes

Abstract The goal of the present work is to model the crystal growth processes mediated by both antisolvent feed and temperature variations through the time evolution of the Particle Size Distribution (PSD). The study is carried out by exploiting two different approaches. In the first approach a population balance equation (PBE) model is devised, where crystal nucleation and growth phenomena are developed taking into account rigorous first principle assumptions. The second approach is based on a phenomenological stochastic formulation leading to a global Fokker–Planck Equation (FPE) governed by a limited number of parameters, describing the time evolution of the probability density function representing the crystal PSD. Validations against experimental data are presented for the NaCl–water–ethanol ternary system. The pros and cons of both approaches are discussed.

A model-based nucleation study of the combined effect of seed properties and cooling rate in cooling crystallization

Abstract Crystallization is a widely used unit operation for the production of pharmaceuticals, fertilizers, and fine chemicals. A commonly used crystallization operational objective is to produce large crystals under minimum nucleation rates. For cooling crystallization there are two key methods for minimizing nucleation, programmed cooling and seeding. In this paper, we evaluate the cooling and seeding methods through the detailed modeling of nucleation phenomena coupled with a population balance and dynamic optimization of this mathematical formulation. Extensive simulation results showed that initial seeding parameters, as proclaimed by others, do have a significant effect on the final product crystal size distribution. It also showed the significance of a combined seeding–cooling approach where a joint cooling and seeding optimization gives superior performance to just optimizing the seed. Importantly, the developed model was highly instrumental in rapidly determining optimal combined seeding–cooling profiles via dynamic model-based optimizations.

Evaluation of the Effect of the Solubility Model on Antisolvent Crystallization Optimization

Abstract Abstract The use of predictive solubility models can be of great use for crystallization modeling, and can decrease the amount of experimental data needed to create a robust crystallization model. In this paper, predictive solubility models such as MOSCED, UNIFAC, NRTL-SAC, and the Jouyban-Acree model are compared against an empirical model for predicted solubility accuracy. The best models are subsequently compared against the empirical model for the antisolvent crystallization of acetaminophen in acetone using water. Two different optimization objective functions are executed for each solubility model to generate corresponding optimal profiles. The effect of these optimal profiles on the predicted crystal properties is evaluated.

A Theoretical Nucleation Study of the Combined Effect of Seeding and Temperature Profile in Cooling Crystallization

Abstract Crystallization is a widely used unit operation used for the production o pharmaceuticals, fertilizers, and fine chemicals. One of the most common types o crystallization is cooling crystallization. In cooling crystallization, the solution is cooled to generate supersaturation causing the formation and growth of crystals. The propertie of the product are dependent on the shape of the supersaturation curve. Nucleation wil occur at high rates if supersaturation increases to excessive levels, this being a undesired outcome. A commonly used crystallization objective is to produce larg crystals with minimum fines production under conditions of low nucleation rates. A typical approach to minimize nucleation known as “programmed cooling” is to utiliz an optimum temperature profile. Recent research works have favored initial seed characteristics over programmed cooling for the production of unimodal crystal siz distributions (CSD). In this paper, a theoretical seed chart is developed for combined seeded-cooling crystallization via fundamental crystallization kinetic modeling o potassium chloride (KCl). Fundamental analysis here shows that joint cooling an seeding optimization of cooling crystallization gives superior performance to jus optimizing the seed, and that the current trend of experimentally optimizing the seed i undermined by model-based optimization approaches.