(714b) Reaction Kinetic Modelling Coupled with Process Safety Calorimetry for Effective Scale-up of an Exothermic Reaction

The present abstract outlines the work carried out for scaling-up an exothermic reaction (ΔT adiabatic = +67ºC) from lab scale (10-20g) to industrial GMP scale (more than 50kg), where the foundations of the scale-up where built upon a fit-for-purpose kinetic model, integrated with calorimetry data and large-scale reactor heat transfer characterization. All the models developed were based on the available experimental results. However, due to time constraints, no further experimental data based on the kinetic modelling findings throughout this work could be generated prior to the GMP campaign.

Primarily, to build the initial model, the proposed mechanism was composed of a competitive lumped reaction system (A+B à P, A+B àImp) and it was developed using experimental data gathered during lab development, namely, routine reaction calorimetry studies and exploring different process conditions such as reaction temperature, reagent, and solvent amount. The reaction model was validated at >100g scale experiments.

Once a good fit between experimental data and mechanistic model was obtained, the heat transfer capability of the large-scale reactor had to be determined. This task was performed by heating and cooling cycles with the same solvent of the reaction and spanning a volume around the reactor occupancy at the target batch scale. By collecting jacket and reactor temperatures throughout the cycles it was possible to determine the heat transfer coefficient of the reactor. Having the cooling capacity of the large-scale reactor determined, it was possible to simulate what-if scenarios at large scale, guide operation conditions during execution of 50kg batches and avoid any uncontrolled temperature peaks during reagent addition and throughout reaction duration. Three scenarios were explored, namely reagent addition at 8 and 20ºC for a 60kg scale batch and 20ºC for a possible future 120kg scale, where the critical aspect was to maintain reaction temperature throughout reagent addition and until reaching the in-process control condition between 5-15ºC. The simulated reaction temperature profiles provided to the production engineers were as follows:

(See pictures)

With this information in hand, it was possible not only to better guide the execution of the batches but also to understand which were the most critical tasks during the reaction (e.g., change in jacket temperature, end of reagent feeding and expected reaction final temperature). The execution of 3 GMP batches (60, 160 and 100kg, respectively) were successfully performed and compared well with the proposed simulated profiles.