Design und Modellierung eines integrierten Bioprozesses zur Produktion natürlicher Aromastoffe
Year: 2005
Type: dissertation
Abstract: Within the scope of this work an integrated bioprocess for the microbial bioconversion of L-phenylalanine to the natural aroma compounds 2-phenylethanol (2-PE) and 2-phenylethyl acetate (2-PEA) was developed and optimised. Organophilic pervaporation was chosen as a procedure for in situ removal of the inhibiting products (2-PE, 2-PEA). Subsequent to characterisation and optimisation of the process, scale up and modelling for description of crossed kinetics were accomplished. A composite membrane made of a permselective layer of PolyOctylMethylSiloxane (POMS) and a microporous support layer of PolyEtherImide (PEI) proved as suitable. In addition to the ability to efficiently remove the aroma compounds, this membrane possesses higher ethanol fluxes than the parallel studied homogeneous membrane. As far as synergistic inhibition goes, due to the aroma compounds in combination with ethanol, this was the decisive factor. The interdependency between sorption and diffusion, as well as the relation from polarity and molar volume, for the studied compounds was made clear by temperature variation of the pervaporation process A significant flux decline of the aroma compounds was observed in the presence of ethanol. Possible causes are competition at free sorption sites, shift of the distribution coefficient and masking of the membrane due to sorption of ethanol molecules. Like ethanol, the use of molasses resulted in a flux decline. For bioconversion two yeasts of Kluyveromyces species (CBS 5670, DSM 5422) were chosen and temperature as well as pH optima (T=40°C, pH 4.50) identified. Experiments on inhibition kinetics for both strains and aroma compounds proved the ester to have a more inhibiting effect than the 2-phenylethanol. Because of the advantage of high temperatures for pervaporative product removal, the influence of increasing temperatures beyond the strain specific optimum was studied. Cultivation at 42.5°C resulted in more than doubled ethanol amounts, when compared to results obtained at a process temperature of 40°C. In combination with aroma compounds and high temperature this resulted in a strong decrease in specific growth rate and decreased product formation rates. Decreasing the ratio of volume to membrane area (V/A) led to a lower accumulation of the aroma compounds in the culture broth and thus to a disturbance in reaction equilibrium. This resulted in increased yields and volumetric product formation rates. A media optimisation with regard to the use of vitamin supplements and different ratios of molasses to dextrose was accomplished. In light of these results it was decided to continue the research without employing the supplements to molasses media. The optimal molasses rates amount to 25% (CBS 5670) and 50% (DSM 5422) respectively. The process, optimised for the separation process and cultivation conditions, was scaled up from a volume of 2L to 200L. Technical limitations of the available bioreactor impeded a reproduction of results observed in laboratory scale. Oxygen dispersion in the medium was insufficient due to reactor and stirrer geometry. This results in an oxygen limitation after a few hours of cultivation. The metabolic changes caused thereby led to deviations on the yields and productivities, when compared to iv Abstract laboratory scale. However, valuable information could be gained over the scale up ability and separation behaviour of the integrated processing procedure. The complete process was modelled with the help of the software package Aspen Custom Modeler™. The model was based on the algebraic description of the intertwined product formation kinetics, associated with product inhibition as well as product removal and enables the dimensioning of the pervaporation plant based on few experiments at laboratory scale. Furthermore statements about the influence of several procedural parameters on the integrated bioprocess can be made and yields as well as productivities can be estimated.
Author: Bianca Maltzahn
Institution: Friedrich-Alexander-Universität Erlangen-Nürnberg
Language: German
Subfield: Biotechnology
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W7018326065
