(528f) Kinetic Modeling of a Non-Homogeneous Metal-Catalyzed Reaction System for Drug Synthesis

Mathematical modelling of chemical reaction kinetics has proven to be a very effective tool for pharmaceutical drug synthesis development, where complex organic reaction networks are often encountered. The chemical and molecular complexity of modern drug molecules has been rapidly increasing, and this requires design and optimization of extensive synthesis routes which presents a challenge to process scientists. First-principles modeling can play an important role in streamlining this process. It helps with analyzing, interpreting and discerning the relationship between various process parameters and reaction profiles with minimal experimentation requirements.

In this presentation, a modelling workflow to understand the Buchwald-Hartwig cross-coupling reactions is described. These reactions are ubiquitous in process chemistry. This is a non-homogeneous, metal-catalyzed reaction system encompassing multiple partially soluble reactants, a biphasic liquid mixture, and homogenous catalytic cycle in the organic phase. Challenges arise from subtle case-to-case variability, with different influential factors like mass transfer limitations, catalyst activation and water sensitivity.

Using this case, the presentation will illustrate the following steps in a typical mechanistic modeling lifecycle: 1) Design and perform kinetic experiments; 2) Recreate experiments in silico; 3) Estimate reaction kinetic parameters; 4) Evaluate model predictions; 5) Postulate new mechanisms, kinetic rate expressions and/or perform additional targeted experimentation until predictions are satisfactory

HPLC measurements were available for key reactants and products from experiments conducted for a range of operating conditions. A mathematical model was developed in gPROMS FormulatedProducts starting from preliminary fitting of primary reaction pathway(s) assuming simple reaction mechanisms and solubility correlations.

Using a systematic parameter estimation approach this was further expanded to a more rigorous model that includes effect of particle size distribution, mass transfer effects and relative concentration effects. The regression of kinetic parameters takes into account uncertainty in measurements and process parameters, which provides a robust model that can be used to make process predictions, and aid scale-up/technical transfer decisions.

The model can be further used to understand the behavior of the system over a wide range of input conditions. This becomes particularly important to make scale up decisions and identify critical operating regimes early in the development of a pharmaceutical compound. Global system analysis was used to explore the design space and identify key factors that influence reaction selectivity and impurity generation. This allows for better understanding of the process and thus provides useful insights to support control strategy, optimization goals, and risk assessment.

Keywords: reaction kinetics, parameter estimation, mechanistic modelling, digital design, catalyst