(614d) A Quality By Digital Design (QbD2) Framework for the Development of Intensified Crystallization Systems

Crystallization is a crucial process for solid material manufacturing serving as both a purification step and a means of controlling critical quality attributes (CQAs). such as crystal size, shape, and polymorphic form.¹ Efficiency in crystallization is prioritized for intensified manufacturing processes with high impact in the manufacturing of solid products for various industries, including energetic, pharmaceutical, food and fine chemicals. Integrating digital twins and digital design frameworks enhances and streamlines this process by providing a virtual representation of the crystallization process, allowing for detailed simulation and optimization across the design space.

In this study, Resveratrol, a dietary supplement, was chosen as the model material. A modular autonomous crystallization system (MACS) with a model-free Quality-by-Control (mfQbC) approach was developed that minimizes the required experiments for developing a process design space and model identification, while utilizing feedback control to enhance the understanding of the system.² An automated design of experiments (ADoE) through an in-house LabVIEW platform called Omnibus streamlines data collection with an automated steady state detection feature. This ADoE including heating-holding-cooling experiments with various cooling rates and seeding amounts to gain insights on crystallization mechanisms, such as primary nucleation, secondary nucleation, the metastable zone width tendencies, and nucleation rate. Direct nucleation control (DNC) and supersaturation control (SSC) experiments were applied to further improve the understanding of attainable regions and design space for the system and to generate data for model identication.¹ Process analytical technology (PAT) tools were used to collect in-situ data that can be translated into crystal size, concentration, and polymorphic form. Based on this information a process model was developed to include the mechanisms occurring, allowing for parameter estimation to then develop a digital twin using two-dimensional population balance modeling (PBM). This digital twin facilitates in silico investigation of the design space as well as the digital design of the crystallization process to achieve the desired CQAs. The results illustrate a systematic quality-by-digital-design (QbD²) framework for the understanding and development of optimized crystallization processes. A one-dimensional population balance model was developed to compare the predictions of crystal behavior to the two-dimensional results. Additionally, this system was transferred from a batch to a continuous crystallization system and the 2D-PBM was adjusted to represent this new configuration.

1. Szilagyi, B., Eren, A., Quon, J. L., Papageorgiou, C. D. & Nagy, Z. K. Application of Model-Free and Model-Based Quality-by-Control (QbC) for the Efficient Design of Pharmaceutical Crystallization Processes. Cryst. Growth Des. 20, 3979–3996 (2020).
2. Eren, A. et al. Development of a Model-Based Quality-by-Control Framework for Crystallization Design. in Computer Aided Chemical Engineering (eds. Kiss, A. A., Zondervan, E., Lakerveld, R. & Özkan, L.) vol. 46 319–324 (Elsevier, 2019).