The role of empirical and mechanistic modelling
The purpose of this work was to build different models to predict the particle size of spray dried powders and assess their usefulness in the design space establishment. Powders were obtained by spray drying solutions of a known pharmaceutical excipient (hypromellose phthalate). The powders were characterized by image analysis (for particle size and circularity), loss on drying (for residual solvent content) and helium pycnometry (for particles density). A full factorial experimental design was performed and PLS regression was used to establish a statistical model. In addition, mechanistic modeling of droplet formation based on hydrodynamic instabilities was also used to estimate the size of particles. Spherical particles, with average particle size between 3 and 9 ?m, were obtained by spray drying solutions in a lab scale unit. Particle density and residual solvent content did not vary significantly between experiments. Statistical and mechanistic approaches were compared. Both statistical and mechanistic models were able to describe the results observed, although the mechanistic model was the most accurate. The mechanistic description of droplet formation was of great assistance to understand and describe the spray drying process.