Statistical Design of Experiment on Contact Secondary Nucleation as a Means of Creating Seed Crystals for Continuous Tubular Crystallizers

Abstract

In the pharmaceutical industry, it is often desired to produce seed crystals with an appropriate narrow size distribution of the desired polymorph. This study describes a system that generates such crystals continuously in a small-scale tubular crystallizer at low supersaturation via contact secondary nucleation. A response surface model was constructed by conducting a statistical design of experiment that models the nucleation rate as a function of contact force, area, and frequency. This model reveals that within a certain range the nucleation rate is linearly related to all three factors in this system. A combination of in-line video analysis and off-line microscope image analysis was used to determine the particle size distribution of seed crystals obtained in this system, and the majority of the crystals were found to be under 20 μm. This seems to be a feature of contact secondary nuclei in general and does not vary significantly with contact force, area, and frequency. Furthermore, the seed crystals generated are of the same polymorph as the parent crystals as a result of the attrition process. This study shows that generating seed crystals with a narrow size distribution using contact secondary nucleation for a continuous tubular crystallizer can be realized at a controlled rate by quantitative variation of certain design parameters.

Overview of Secondary Nucleation: From Fundamentals to Application

“…Under the action of reasonable force, the contact secondary nucleation caused by the microattrition of the parent crystal is considered as the main mechanism to control the crystal form; therefore, the secondary nucleus often has the same polymorphism as the mother crystal, thereby achieving control of polymorphism. 58 Seeding is usually used as a source of the secondary nucleus, which controls nucleation by contact with the surface of the seed or contact propagation of nuclei swept from the surface of the seed by hydrodynamic forces. 38 Miura et al 183 were able to form the desired racemic sulfonic acid metastable δ-polymorph by adding the previously formed δ-polymorph single crystal to the supersaturated solution.…”

“…This is an ideal choice for controlling polymorphism, indicating the inherent stability of the steady-state polymorphism of the MSMPR system. 185 Cui et al 58 described a system for continuously generating specific crystal forms under a low supersaturation state by contacting secondary nucleation in a small tubular crystallizer and proposed that due to the attrition process the generated crystals are the same as the parent crystal. Lai et al 185 used enantiotropic p-aminobenzoic acid at either the α or β polymorph as a model to expand their understanding of polymorphism.…”

“…Therefore, when we deal with the polymorphism system, one can employ such a secondary nucleation event to produce the same polymorphism as the seed crystal, thus achieving control of the polymorphism outcome during the crystallization process. 58 However, some studies 59,60 have shown that the contact secondary nucleus comes from the absorption solute layer between the crystal surface and the bulk solution. Shi et al 61 simply summarized the formation of contact secondary nucleation in two cases: (i) tiny crystal fragments detach from the contact surface or (ii) release of sufficiently large embryos into saturated solution by a hypothetical growth layer.…”

Automation Technologies to Enable Data-Rich Experimentation: Beyond Design of Experiments for Process Modeling in Late-Stage Process Development

“…Using numerical techniques to guide the optimization procedure offers significant efficiency gains over iterative, experimental approaches alone. Optimization of experimental factors through response surface modeling is not uncommon; − however, the unique ability to include time in the optimization via the DRSM model expands the optimization problem set to include time-dependent responses, such as productivity, cycle time, and stability.…”

Iterative model‐based experimental design for spherical agglomeration processes

“…Once an appropriate mathematical model has been identified, the model parameters need to be estimated accurately for in‐silico predictions of process outcomes. Traditionally, experimental design for model parameter estimation have been carried out using a statistical design of experiments (DOE) 45 including full factorial design, and fractional factorial design, among others. However, increased number of experiments is typically necessary to get precise estimates of model parameters using traditional DOE methodologies.…”