(420e) A Novel Approach for Modeling Pharmaceutical Tablet Coating Using Coupled DEM-CFD Modeling

Oral drug delivery via tablets is highly favored across numerous therapeutic areas for its convenience, precise dosing, stability, and patient adherence. Tablets often receive a thin film coating on their surface. In most cases, the coating is applied simply for cosmetic reasons, e.g. branding, blinding, etc. A small subset of tablet formulations, however, require a functional coating. An example of functional coatings is the one applied to prevent drug dissolution in the stomach and instead release it in the small intestine (i.e. enteric coating). This coating is essential for protecting the drug from gastric acid degradation or inactivation, which is often a requisite for oral delivery of macromolecules.

The coating process is performed on a rotating perforated cylindrical pan with conical ends, often with baffles designed to tumble the tablets therein. The coating formulation is then sprayed onto a small area of the tablet bed via an atomizer. The spray droplets that deposit onto the tablet surface as it passes through the spray zone are then dried to form a coating film by the warm air that is passed through the pan’s perforations. The process is continued until an established endpoint, typically an average tablet weight gain, is achieved. For nonfunctional (i.e. cosmetics) coating, an average weight gain of 4-5% is usually targeted. For enteric coating, the target weight gain can be as large as 15-20%.

Tablet weight gain is typically assessed by first taking the average weight of 10-20 (uncoated) tablets before starting the coating process. Once the coating begins, the same number of tablets as that initially measured are drawn from the pan. Their weight is then measured. This process is repeated at a fixed time interval until the target weight gain is achieved. In the case of enteric coating, the weight of individual tablets representing a small portion of the batch may be measured as well to ensure that the inter-tablet coating uniformity (i.e. the amount of coating deposited between one tablet and the next) is adequate. While this approach to quality control may seem primitive, it is frequently deemed sufficient because measuring the weight of all tablets in a batch is practically impossible, especially when the batch size is large (i.e. hundreds of kilograms). Accordingly, an alternate method that better quantifies the process outcomes, such as numerical methods that incorporate all key process physics and thus can predict the inter-tablet coating uniformity more accurately than the above experimental approach, can be extremely valuable for process development and validation.

This study presents the development of a novel DEM model using Ansys Rocky to predict inter-tablet coating uniformity. The model incorporates the actual tablet shape (instead of using a “glued spheres” approach). A hysteretic linear spring model is used to compute normal contact forces, and a linear spring Coulomb limit model calculates tangential forces. Spray droplets are modeled as DEM particles and tracked within the DEM framework, allowing the local spray properties (i.e. droplet concentrations at a given location, their velocities, and size distributions) to be captured and thus simulating the spray dynamics in the coating process. Contact detection between spray particles is omitted to save computational resources. Additionally, upon collision with a tablet, a droplet is immediately removed from the simulation, and its mass is added to the liquid film mass on that tablet. In addition, the size of each tablet is also increased in the simulation to account for the increasing coating mass deposited on them, which, in the case of enteric coating, can be significant when compared to the uncoated tablet weight itself.

To better model the spray-tablet interactions, the mass, momentum, and energy transfer between the spray droplets and the underlying flow field are also modeled by coupling Ansys Rocky and Ansys Fluent solvers. A droplet evaporation model is implemented by computing the instantaneous local evaporation rate based on the difference in water vapor concentration between the particle surface and its surroundings. As the water evaporates, the remaining solute fraction is integrated into the coating deposit on the tablet surface from the previous time step. As explained above, this additional mass (and volume) is then used to increase the mass and size of the tablets in the simulation. Finally, a liquid bridge model is also incorporated to account for liquid transfer between tablets that may occur during tablet-tablet collisions, as well as the effect of viscous lubrication and capillary forces on the particle bed dynamics.

We demonstrate the effects of adjusting process parameters such as droplet size, spray rate, drum speed, and temperature on process outcomes such as the time to reach the coating endpoint, inter-tablet coating uniformity, and tablet bed dynamics. Model results are also compared to experimental data from bench-scale tablet coating runs.