Process development for plant-based extract production

In this thesis a structured and standardized approach for the process design of plant-based extractions has been developed. This approach is founded on laboratory experiments performed in a standard apparatus to determine extraction kinetics for the best solvent and the targeted solute with regard to the solubility, selectivity as well as factors like regulatory constraints and consumer demands. As a benchmark, two different raw materials, pepper berries and vanilla beans, with different physical properties of the target compounds and the raw materials have been examined to cover a wide range of typical botanical variable feeds. Furthermore, mass balances have been set up for water, solvent, raw material and target compounds on a laboratory scale. A rigorous model supported by botanical investigations has been developed to predict percolation results. After determining the desorption isotherm the model has been validated with short residence time percolation experiments. With the aid of this validated model percolation experiments with long residence times can be predicted and different process scenarios calculated. A combination of design of experiments and rigorous modelling in combination with a standard equipment and experimental model parameter determination provides a fast and robust experimental design on the one hand and the precondition for a process parameter data-based learning curve on the basis of rigorous modelling in the mid- or long-term on the other hand.