271453 Mixture Component Prediction Using Iterative Optimization Technology

Tuesday, October 30, 2012: 10:30 AM
Allegheny II (Westin )
Koji Muteki1, Daniel O. Blackwood1, Kyle R. Leeman1, George L. Reid1, Yong Zhou1, Yang A. Liu1, Brent Maranzano1, Paul R. Gerst1, Howard W. Ward1, Andreas AM. Mühlenfeld2 and Matthias Danner2, (1)Pfizer Worldwide Research and Development, Groton, CT, (2)Pfizer GMS, Freiburg, Germany

Spectroscopy techniques such as near infrared (NIR), mid Infrared (MIR) and ultraviolet (UV) are commonly used to monitor various process operations such as blending, mixing and chemical reactions in pharmaceutical development. The calibration required for a quantitative measurement can be a labor-intensive and time-consuming to implement. The objective of this study is to predict/monitor/estimate the mixture component ratio without making calibration standards in advance (e.g., minimum effort) while keeping similar prediction ability as with calibration standards. Reference methods such as HPLC/GC are used just for the validation purpose.

Iterative Optimization Technology (IOT) involves the use of an optimization computation using only pure component spectra collected prior to a batch and mixture component spectra collected during a batch. The proposed approach has been successfully applied to many practical projects (solvent mixtures, chemical reactions, feed-frame process, blending process). It is expected that the use of IOT will enhance the adoption of PAT as a routine analytical tool in project labs, especially for process understanding and early development activities.


Extended Abstract: File Not Uploaded