Multivariate curve resolution for kinetic modeling and scale-up prediction

Abstract

An imine synthesis was investigated in a nearly isothermal oscillating segmented flow microreactor at different temperatures using non-invasive Raman spectroscopy. Multivariate curve resolution provided a calibration-free approach for obtaining kinetic parameters. The two different multivariate curve resolution approaches, soft and hard modeling, were applied and contrasted, leading to similar results. Taking heat and mass balance into account, the proposed kinetic model was applied for a model-based scale-up prediction. Finally, the reaction was performed in a 0.5 L semi-batch reactor, followed by in-line Raman spectroscopy and off-line gas chromatography analysis. The successful scale-up was demonstrated with a good agreement between measured and predicted concentration profiles. Highlights • Oscillation segmented flow reactor with inline Raman spectroscopy. • Multivariate Curve Resolution with hard and soft constraints. • High quality kinetic model for scale-up predictions. Graphical abstract

Simultaneous reaction- and analytical model building using dynamic flow experiments to accelerate process development

“…Numerous platforms have been developed to establish and parameterize kinetic models using data generated in ow, but these either focus on a specic reaction type (not applicable to general synthetic organic chemistry), 26,27 use relatively simple data-processing methods (pre-developed chromatography methods or spectroscopy considering only a single peak), [28][29][30][31] or require a high degree of manual data processing. [32][33][34][35][36] The aforementioned kinetic model parameterization platforms generally rely on single data points, which are measured once the reactor has reached steady state. Utilizing dynamic experiments can drastically accelerate the collection of dense datasets.…”