(558g) On-Line FTIR for Accurate Fundamental Kinetic Analysis, Real-Time Process Monitoring, and Process Controls Justification in Pharmaceutical Manufacturing

Authors: 
Marek, J. C., AbbVie, Inc.
Dunn, T. B., AbbVie

On-Line FTIR for Accurate Fundamental Kinetic Analysis, Real-Time Process Monitoring, and Process Controls Justification in Pharmaceutical Manufacturing

Eric G. Moschetta,* James C. Marek, and Travis B. Dunn

AbbVie, Inc. Process R&D, 1401 Sheridan Road North Chicago, IL 60064

*corresponding author: eric.moschetta@abbvie.com

Fundamental kinetic analysis of complex reactions is critical to process understanding and development in the pharmaceutical industry. With aggressive timelines for active projects, it is imperative to design and execute data-rich experiments to glean as much kinetic information in as few experiments as possible. However, existing analytical techniques may not be fully developed for complete kinetic analysis, nor may they be amenable to collecting many data points in a single experiment. For example, reaction samples that require labor-intensive workup or processing to analyze quantitatively do not lend themselves well to data-rich experimentation because of the significant burden on the development scientist and the available instrumentation. An alternative approach is to implement on-line process analytical technology (PAT) tools for monitoring the concentrations of the relevant chemical species participating in the reactions. On-line measurements are amenable to collecting many data points over the course of an experiment because the instruments often have rapid sampling frequencies. PAT tools that rely on spectroscopy offer the additional benefit of structural characterization and can distinguish between different chemical species. Careful identification and assignment of characteristic peaks and bands from the on-line signals can provide time-dependent concentration data simultaneously for multiple chemical species. These data can be used not only to build a robust chemometric kinetic model for the process, but for simple identification of logical process endpoints or even to monitor critical quality attributes (CQAs) in real time. As such, implementing on-line PAT tools to improve process understanding at a fundamental level and streamline the workflow in the laboratory is critical to developing robust manufacturing processes in the pharmaceutical industry.

We present a business case in which we implemented an on-line FTIR probe for in situ operation in data-rich kinetic experiments. Our business objective was to justify the process controls for the rate of addition of a reagent into a chemical reactor such that the starting material in the reactor was always consumed to the justified level. We hypothesized that a kinetic model describing the decomposition of the starting material would provide excellent justification for the process. However, the kinetic analysis was complicated by two factors: (1) the reagent being dosed to the reactor could further react with the desired product to form an impurity and (2) the current HPLC method required a time- and labor-intensive derivatization process, limiting the amount of data points that could be collected and analyzed in any given experiment. All of our relevant kinetic species contained different functional groups bound to a carbonyl group, so we pursued on-line FTIR as a possible PAT tool to acquire data. We will show how the key experimental observations were more easily elucidated by the on-line IR than the off-line HPLC method and how we used these observations to design and execute the minimum number of data-rich experiments to build a robust kinetic model of the process. Specifically, we will explain how we were able to observe a subtle reaction effect using the on-line FTIR that proved our working hypothesis of the reaction mechanism was wrong. Finally, we show how our chemometric model can be used to monitor pilot plant scale processes in real time based on the fundamental principles developed in the laboratory. This level of understanding allows for predictive power in scalable processes and provides excellent details for process control justification.

The design, study conduct, and financial support for this research was provided by AbbVie. AbbVie participated in the interpretation of data, writing, reviewing, and approving the publication. All authors are AbbVie employees.