(668b) Development of a Process Signature for Fluid Bed Drying Using Nir Spectroscopy

Near infrared spectroscopy can be used to determine the product's moisture content real-time in a fluid bed dryer. Creating a robust moisture model requires extensive sampling to establish a correlation between the NIR spectra and the moisture content. In addition to moisture information, the NIR spectra also capture changes in the physical properties of the material. Therefore, expanding the scope from a moisture analysis to a more inclusive process signature determination can aid both formulation and process development and scale-up. Qualitative analysis can be used to establish a process signature or trend that can be used to compare various lots of different formulations and at different scales. The process signature was calculated using a principal component analysis. The spectroscopic data was analyzed and viewed using the scores vs. scores plot. Using the same principal components, the scores can be calculated for independent batches real-time. The deviation from previous runs can then be calculated. Monitoring the process signature during scale-up insured the large scale drying process is consistent with the small-scale work. By monitoring the NIR spectra through drying runs at the 10 kg, 60 kg, and 300 kg, a representative process signature was obtained and employed to ensure consistency between development and manufacturing scale batches. The information was then used to set the manufacturing process parameters. Furthermore, by continuously monitoring the process signature at the manufacturing scale, it is possible to identify batches that are drying differently and compensate for that by altering the drying time or temperature. Obtaining a process signature during development can facilitate process scale-up and reduce batch-to-batch variability at the manufacturing scale.