Moving Horizon Parameter Estimation for an Enzyme-Catalyzed Transamination Reaction with Integrated Product Removal

Chiral amines are key building blocks in producing active pharmaceutical ingredients and can be synthesized via the amine transaminase pathway under mild conditions with eco-friendly solvents. For model-based design and control, mathematical models can be utilized. This requires reliable parameterization of the model. In this study, a model for the production of (S)-(3 methoxyphenyl)-ethylamine is presented. Parameter identifiability is analyzed using the method of profile likelihoods. The parameter identifiability is then tested in two different batch operation scenarios by using moving horizon estimation, which can be applied for on-line model adaption and control purposes. In a batch process, identifiability decreases over time. A repetitive batch strategy improves parameter identifiability.