(740g) Wurster Coating: Prediction of Undesired Agglomeration

Wurster fluid bed coating is an effective way of manufacturing pharmaceutical pellets requiring specific characteristics such as high drug loading or enteric protection. Optimization of the coating process can be demanding because it necessitates many time-consuming pilot-scale experiments, the scope of which is limited by the cost of active pharmaceutic ingredients. To reduce the need for experiments and prevent undesired agglomeration, a multi-scale model has been developed.

The model utilizes statistically evaluated particle dynamics as well as heat and mass transfer resolved for both device and particles. The model was validated against existing production data, and the predicted agglomeration probability was in agreement with the observed undesired agglomeration or successful coating regime.

Wurster fluid bed coating is one of the most versatile coating methods [1]. It offers fast heat and mass transfer, as well as temperature homogeneity. In contrast to conventional fluidized bed coating, a coating solution is sprayed from a nozzle placed bottom-up in the center of a cylindrical tube (draft tube). Particles are driven through the tube and periodically move through the spraying zone, where they collide with small droplets of the coating solution. The circulation of the particles in this way ensures good mixing and provides better control of particle movement and the resultant coating properties.

In comparison to other coating methods, the main disadvantage of Wurster is the tedious setup and process optimization [1]. Currently, manufacturers choose this method mainly for special applications like formulation of sustained or enteric release. Coloring, odor, or taste masking, where uniform film thickness and coating porosity are less important, can be done in a conventional top-spray fluidized bed.

In this work, we mainly focus on the coating of small particles for a long period, which is the usual case for the deposition of an API. The long process time assures the low inter-particle coating variability [2].

However, agglomeration is here the main concern. Agglomerated pellets must be separated before capsule filling, reducing the yield. If agglomerated pellets again disjoin before or during separation, the resulting inter-particle variability (API dosage) will be higher. If combined with sustained release or enteric protection, the whole formulation might be ineffective.

Agglomeration has several causes. An insufficient drying rate can lead to excessive wetness and result in the coating solution acting as a binder. Conversely, excessive drying or low air humidity can lead dry particles to generate electrostatic charges and, thereby, stick together. To prevent agglomeration, it is necessary to keep the wetness of the particles within a narrow threshold range while maintaining homogeneous coating distribution. However, the coating distribution and particle wetness can only be controlled indirectly by setting operating parameters, such as temperature, spraying rate, draft tube position, volumetric airflow rate, and humidity.

Agglomeration can be viewed in terms of probability as a product of geometrical and the physical collision success factors [3]. The geometrical factor is based on surface coverage, and the physical factor is based on coalescence conditions derived from Stokes number analysis [4].

CFD-DEM simulation can accurately predict the particulate flow behavior. Using CFD-DEM directly coupled with droplet deposition, drying, and particle liquid bridges interactions is so strongly limited by the number of particles that any application to a real pharmaceutical process where several millions of particles are present seems impossible. However, a promising approach is a statistical evaluation of the repetitive patterns in particle circulation [5].

We follow this idea and present a model that uses statistically evaluated particle dynamics. Additionally, at the onset of the agglomeration, the inlet spray rate, air humidity, and inlet air temperature have almost no impact on the particle dynamics. And the drying kinetics can be analyzed separately with lower computational effort. On the device-scale, the model predicts the temperature, moisture content, and average coating thickness. On the particle-scale, model computes particle dynamics via CFD-DEM. On the micro-scale, the model estimates the coating deposition, drying kinetics and change of its physical properties. By coupling models at all length-scales, including only statistical analysis of particle movement and Stokes coalescence condition, it is possible to evaluate particle agglomeration probability and use it as a tool for process optimization and possibly for transfer between different production sites and equipment.

We have analyzed data from 13 production records. All experiments were carried out in Glatt GPCG 2 fluid bed equipped with a 7'' conical Wurster chamber and a cylindrical expansion chamber. Spraying solution consisted of ethanol or isopropanol together with a full pharmaceutical formulation including the API and polymer. And the predicted agglomeration probability was in agreement with the observed undesired agglomeration or successful coating regime.

[1] P. Hadfield, Processing and Equipment Considerations for Aqueous Coatings, (4th ed.), CRC Press, New York (2016), pp. 49-82

[2] X.X. Cheng, R. Turton, The prediction of variability occurring in fluidized bed coating equipment. II. The role of nonuniform particle coverage as particles pass through the spray zone, Pharm. Dev. Technol., 5 (2000), pp. 323-332

[3] P. Rajniak, F. Stepanek, K. Dhanasekharan, R. Fan, C. Mancinelli, R. Chern, A combined experimental and computational study of wet granulation in a Wurster fluid bed granulator, Powder Technol., 189 (2009), pp. 190-201

[4] L.X. Liu, J.D. Litster, S.M. Iveson, B.J. Ennis, Coalescence of deformable granules in wet granulation processes, AICHE J., 46 (2000), pp. 529-539

[5] T. Lichtenegger, S. Pirker, Recurrence CFD – a novel approach to simulate multiphase flows with strongly separated time scales, Chem. Eng. Sci., 153 (2016), pp. 394-410