Comparison of distillation strategies, optimization and scale-up for industrial processes based on mechanistic modelling

corresponding

RUI PINA CAMPOS, FILIPE ATAÍDE, ANTÓNIO HENRIQUES, JOÃO SARDINHA
Hovione Farmaciência SA, Lisboa, Portugal

Abstract

API production/chemistry processes commonly include distillation operations for several purposes and at different stages and steps of the process. While distillation is typically not considered as a critical or problematic operation, their study and mechanistic modelling can bring real gains to the process, namely by saving development manufacturing times, solvent volumes and reducing the number of required lab experiments to optimize the process conditions. The application of Hovione’s methodology for distillation models was conducted for two different industrial projects resulting in specific distillation/solvent-swap models for the API processes. In the case studies presented, different distillation strategies to achieve a target final composition are compared. The volumes of fresh solvent needed is also reduced, without impacting the process outcome. Modelling studies to support the scale-up from laboratory to pilot plant scale, resulting in optimized distillation conditions and reduction of operation time, are also presented.


INTRODUCTION

Process development and manufacturing in the Pharmaceutical Industry comprises several unit operations that need to be studied and optimized along the drug development lifecycle. The chemical development is complex and challenging as it involves large effort from multi-disciplinary teams, with tight timelines and limited resources for experimentation.(1) Distillation is one of the most common unit operations of the API’s chemical processes and it can be applied for several purposes like solvent-swaps, concentrations, crystallizations or even purifications (2-4).

 

At a time where Development by Design is so important in shaping the future of the Pharmaceutical industry, it is essential to adopt data-driven methodologies and (mechanistic) models to improve chemical processes (6). Unlike empirical or statistical models, mechanistic models consider and relate the mathematical descriptions of the chemical and physical processes, reflecting physical laws (mass balance, energy balance and heat transfer relationships) (6-7). As illustrated in Figure 1, the models provide accurate answers to import ...