(164f) Implementation of Quality-by-Design Principles: Modeling of Spray Tablet Coating Using a CFD Approach

Implementation
of Quality-by-Design Principles: Modeling of Spray Tablet Coating Using a CFD
Approach

Ariel
R. Muliadi and Paul E. Sojka

Maurice
J. Zucrow Laboratories

Purdue University School of Mechanical Engineering

Inter-tablet coating non-uniformities can arise from
non-uniformities in sprays [1].  A recent study by Muliadi and Sojka [2] confirms that typical sprays used in pharmaceutical tablet coating have non-uniform distributions of size, velocity, and mass flux.  In another study, Muliadi and Sojka [3] show that spray drop, velocity, and mass flux distribution can vary with drum coater operational parameters.  As an example, they found that supplying drying air to the pan-coater results in up to a 6 m/s increase in drop velocity, but has mixed effects on drop size.  Furthermore, when the spray is composed of mostly small drops (D₃₂<12 μm), supplying drying air to the drum leads to an increase in D₃₂.  The reverse is observed for sprays with D₃₂>12 μm.  These two observations are most evident when operating at low liquid supply rate (70 g/min), suggesting that they likely arise from drop evaporation.  Finally, adding drum rotation to the process generally leads to an increase in drop size and a decrease in drop velocity.  These two effects are greatest when the nozzle is operated as an air-assist atomizer, most probably because of the production of smaller droplets. 

Muliadi and Sojka's results [3] suggest that spray evolution in a pan-tablet coater is influenced by complex transport processes.  In order to determine how these processes interact with one another and with the sprays, a computational model was set up in FLUENT.  The model considers a Eulerian method to simulate spray gas-phase and drying air flow field inside a drum coater, as well as a Lagrangian approach to compute spray drop trajectories.  Actual spatially resolved spray data were used as inputs to the Lagrangian model. 

Computed data agree well, to within
experimental uncertainty with the experimental data presented in [3].  Other notable observations include the following:

o      
The dumbbell-shaped spray pattern remains regardless of the drum
operating conditions.  Further, since the drying air flow path is not perfectly
aligned with the gun's axis, it tends to push the sprays downward; leading to
spray patterns that resemble the letter ?C'.  This can likely be avoided by
adjusting the gun's orientation so that its axis is perfectly aligned with the
drying air flow path, or by placing the gun at a location not passed by the
drying air flow. 

o      
Drying air flow reduces spray drop number density by promoting
evaporation.  At the same time, it also acts as co-flow for the spray drops,
therefore increasing the droplets' mean axial velocity and reduces the spray
extent.  The combination of these two effects leads to an increase in spray
local volume flux.  Drop evaporation also leads to an increase in average drop
size (D₃₂). 

o      
Drum rotations decrease spray drop number density by transporting
small drops away from the spray region.  This also increases drop size,
particularly near the spray center, where the small drops tend to be
clustered.  The decrease in spray drop number density following the increase in
drum rotation is less significant for processes with drying air supplied.  This
suggests that drying air flow helps the drops oppose the effects of drum
rotations by providing them with additional axial momentum.  In addition, drum
rotation also appears to slightly increase the spray extent, which in turn
increases the distance between the two volume flux peaks.  It therefore essentially
makes the spray less uniform.

o      
Adding drum rotations to the process makes the drying air flow
path less aligned to the gun's axis.  As a result, the drying air will provide
less axial momentum for the drops.  This leads to a decrease in drop velocity. 
Drum rotation also transports drops away from the spray region.  This leads to
a decrease in spray drop number density.  The decrease in drop velocity,
combined with the decrease in spray drop number density, in turn results in a
decrease in local spray volume flux. 

References:

1. Mowery, M. D., Sing, R., Kirsch, J., Razaghi, A., Bechard, S., and Reed, R. A.,
2002, Journal of Pharmaceutical and Biomedical Analysis, 28(5), pp.
935-943.

2.
Muliadi, A. R., and Sojka, P. E., 2009,  11th International Conference on
Liquid Atomization and Spray Systems (ICLASS) Vail, Colorado.

3.
Muliadi, A. R., and Sojka, P. E., 2009,  2009 ASME International Mechanical
Engineering Congress & ExpositionLake Buena Vista, Florida.