Organic synthesis is a critical step in producing any Active Pharmaceutical Ingredient (API): it encompasses a series of organic (mostly catalytic) reactions and separations, whose purpose is to arrive at a solution of a target molecule, which must then undergo crystallisation for purification, followed by secondary (downstream) processing towards the final (commercially available) formulation. Consequently, studying and optimising industrial-scale API production requires reliable process unit (especially reactors and separators) and system descriptions, to track key chemical (esp. impurity) concentrations. The present paper describes a Digital Twin (DT) which is currently under development for a particular API target: its goal is to evaluate multicomponent adsorption of Volatile Organic Compound (VOC) emissions from licensed salbutamol synthesis, via industrial FTIR data for mixtures of dichloromethane, chloroform, toluene, methanol and ethanol. A combination of first-principles (e.g. PDE-based adsorption) and data-driven (e.g. PCA) methodologies is essential in order to ensure techno-economically optimal performance.

• University of Edinburgh
• GlaxoSmithKline

• Modeling

1. Digital Twin (DT)
2. Hybrid Modelling
3. modeling
4. Pharmaceutical Manufacturing
5. Research Article