Design and scale-up procedures for granulation in the nucleation phase have been established for a Fielder mixer granulator. The procedure is designed to maintain product attributes such as high density and narrow size distribution, across scale for use in the pharmaceutical industry. The dimensionless spray flux, proposed by Litster et al., 2001a, and the drop penetration model for loosely packed beds proposed by Iveson et al., 2001a were the key dimensionless parameters used to ensure drop controlled granule growth. A high impeller speed was used to ensure optimal contact of the powder with the liquid binder. Results from Hapgood, 2000 were used to illustrate the design procedure.
The nucleation design procedure was scaled from a 25L to a 300L Fielder mixer granulator. Impeller speed was scaled maintaining a constant Froude number to ensure binder dispersion and powder mixing were maintained across scale. Adjustment of the number of nozzles used to disperse the liquid was necessary. The effects of the chopper were ignored.
The effects of binder viscosity on granulation in the nucleation phases were discussed. Highly viscous liquid binders have a long penetration time, which pushes granulation out of the drop controlled region into the mechanical dispersion region, where granulation is controlled by mechanical mixing and agitation (Hapgood, 2000). Unless the drop penetration time of the liquid binder can be reduced to fall within the drop controlled region, scaling is pointless since nucleation by mechanical dispersion is uncontrolled and causes segregation.
The design and scale-up issues of growth and consolidation in Fielder mixer granulators were addressed. A relation of the change in porosity with time must be established in the small-scale granulator. Consolidation depends on the granule deformation and operating conditions. To maintain constant initial and minimum porosity across scale, the collisional velocity, which is related to the tip speed, must be kept constant. The Stokes deformation number is a function of the collisional velocity and granule yield strength. Impeller speed is scaled using a constant Stokes deformation number. This will maintain a constant tip speed across scale. Design procedures were based on reports by Mackaplow et al., 2000, and Schaefer et al., 1986.
To accommodate for the different impeller speeds required during different phases of granulation, a “batch program” is proposed. The program involves using the impeller speed scaled by Froude number during the nucleation phase and the Stokes impeller speed for consolidation. Validation of the program is required since the time frame for each phase is dependent on the formulation and process properties and cannot be determined without prior knowledge of the system.
A sensitivity analysis was performed to determine the validation of the assumptions made in the design procedure. True effects of the parameters are seen in the product size, density and size distribution. The results from the sensitivity analysis implied that the nucleation design procedure for was reliable for nucleation in the drop controlled region.