Experimental Study of Fluidized Bed Co-Granulation of Two Active Pharmaceutical Ingredients: An Industrial Scale-Up Perspective

Abstract

This article presents our efforts toward the development and scale-up of a fluidized bed process for co-granulation of two highly water-soluble active pharmaceutical ingredients (APIs) with a polymeric binder. The granulations were produced at three scales using Glatt GPCG3 (small), GPCG15 (pilot), and GPCG120 (commercial) fluid bed columns. The effects of binder concentration, atomization air pressure, and inlet air temperature were first investigated at the small scale. Based on the process knowledge generated at the small scale, various granulations were produced at the pilot scale to determine the effects of solution spray rate. We have demonstrated that, for a given binder solution, the droplet size distribution can be maintained similar by adjusting the atomization pressure when different spray rates were used upon scale-up. Superficial air velocity and moisture level in the powder bed appear to be the other key response parameters, which are, in turn, governed by various process and design parameters as well as formulation properties. This article demonstrates that the process and product quality remains invariant when the aforementioned key response variables are maintained during scale-up through some simple scale-up rules.

Keywords:

• atomization
• droplet size
• fluidized bed
• granulation
• pharmaceuticals
• process scale-up

Acknowledgments

The authors would like to thank Ashkan Kamali, Monica Tijerina, Soumojeet Ghosh, James Zega, and Scott Reynolds for their collaboration and support during this work. E. B. would like to extend thanks to Anagha Bhakay at NJIT's Chemical Engineering Department for proofreading the manuscript and offering constructive criticism.

Notes

^a Based on the theoretical assumption that water was completely removed during the granulation.

^b Residual water is targeted at ≤1%.

^a Taken as a descriptor of the fluidized bed granulation scale.

^b The nozzles at the 10 L and 45 L scales are similar single-headed two-fluid nozzles (Schlick Model 941, Form 7–1); the nozzle at the 420 L scale is a three-headed two-fluid nozzle (Schlick Model 937).

^a Binder concentration in the aqueous solution.

^a Binder solution concentration.

^b Duration of spray with binder spray rate in step 1 divided by the total spraying time.

^c Minimum and maximum of the actual values that were set by the inlet damper.

^d Calculated based on the amount of binder solution added via the peristaltic pump.

^a At the end of the spraying phase.