Library

Notes: Abstract The scale-down procedure can be used to optimize and scale up fermentation processes. The first step in this procedure, a theoretical analysis of the process at a large scale, must give information about the regime, or bottle necks, ruling the process. In order to verify the theoretical results the process analysis has been applied to the fed-batch baker's yeast production at a laboratory scale. The results of this analysis are compared with results from fed-batch experiments. It was concluded that if only one mechanism is ruling the process, for instance mass transfer, the results of the analysis are quite clear. If more than one mechanism is important, for example mass transfer and liquid mixing, additional knowledge is needed to predict the behaviour of the process. Concerning the baker's yeast production, it was concluded that if oxygen limitation occurs, liquid mixing is of little importance.

Notes: Abstract Immobilization is a method of avoiding wash-out of biocatalyst from a reactor system. For the modelling of these biocatalysts slab, cylinder, sphere and biofilm geometries are frequently used. A biofilm particle consists of an inert core which is used as a carrier for a layer that contains the enzymes or micro-organisms. This paper deals with the modelling and effectiveness factor calculations for such a biofilm particle and a general model for an immobilized, non-growing biocatalyst is presented. The model includes internal and external mass transfer resistance, the partitioning effect and inhibition or reversible reaction kinetics. Due to the non-linear reaction rate equations of the Michaelis Menten type, numerical techniques must be used for the solution of the combined diffusion reaction equation and calculation of the effectiveness factor. In this work we have used two different methods, orthogonal collocation and a method based on Runge-Kutta integration. Comparable use of CPU-time was found for these methods, but numerical stability and accuracy favour the Runge-Kutta method. In the case of Michaelis Menten kinetics (irreversible and without inhibition effects), an analytical expression for an approximate solution is presented. This method, which has an acceptable accuracy, takes far less CPU-time than the fore-mentioned numerical techniques.

Notes: Abstract The productivity of a fermentation is proportional to the biomass concentration. The productivity can therefore be increased by retention of the cells in the fermentor. In this study microfiltration was used for cell retention in a fermentation of glucose to ethanol by baker's yeast. Compared to a system without cell retention the productivity could be increased 12-fold to 55 kg/m3 h at a biomass concentration of 135 kg/m3. Maximal ethanol concentrations of 76 kg/m3 were obtained at conditions of growth. At zero growth conditions in the integrated system the ethanol concentration could be increased to about 115 kg/m3, and could be produced for at least 10 hours. The fermentation results in the integrated system could be described reasonably well with a mathematical model based on a different linear inhibition kinetics for growth and substrate consumption.

Notes: Abstract Extraction can successfully be used for in-situ alcohol recovery in butanol fermentations to increase the substrate conversion. An advantage of extraction over other recovery methods may be the high capacity of the solvent and the high selectivity of the alcohol/water separation. Extraction, however, is a comprehensive operation, and the design of an extraction apparatus can be complex. The aim of this study is to assess the practical applicability of liquid-liquid extraction and membrane solvent extraction in butanol fermentations. In this view various aspects of extraction processes were investigated. Thirty-six chemicals were tested for the distribution coefficient for butanol, the selectivity of alcohol/water separation and the toxicity towards Clostridia. Convenient extractants were found in the group of esters with high molar mass. Liquid-liquid extraction was carried out in a stirred fermenter and a spray column. The formation of emulsions and the fouling of the solvent in a fermentation broth causes problems with the operation of this type of equipment. With membrane solvent extraction, in which the solvent is separated from the broth by a membrane, a dispersion-free extraction is possible, leading to an easy operation of the equipment. In this case the mass transfer in the membrane becomes important. With membrane solvent extraction the development of a process is emphasized in which the extraction characteristics of the solvent are combined with the property of silicone rubber membranes to separate butanol from water. In the case of apolar solvents with a high molar mass, the characteristics of the membrane process are determined completely by the solvent. In the case of polar solvents (e.g. ethylene glycol), the permselectivity of the membrane can profitably be used. This concept leads to a novel type of extraction process in which alcohol is extracted with a water-soluble solvent via a hydrophobic semipermeable membrane. This extraction process has been investigated for the recovery of butanol and ethanol from water. A major drawback in all processes with membrane solvent extraction was the permeation of part of the solvent to the aqueous phase. The extraction processes were coupled to batch, fed batch and continuous butanol fermentations to affirm the applicability of the recovery techniques in the actual process. In the batch and fed batch fermentations a three-fold increase in the substrate consumption could be achieved, in the continuous fermentation about 30% increase.

Notes: Abstract In large-scale bioreactors gradients often occur as a result of non-ideal mixing. This phenomenon complicates design and control of large-scale bioreactors. Gradients in the oxygen concentration can be modeled with a two-compartment model of the liquid phase. Application of this model had been suggested for the control of the dissolved oxygen concentration with a batch gluconic acid fermentation process as the model system. The control system consists of a controller, an observer and a parameter estimator. In this work, the controller design is reconsidered and, in simulation experiments, the performance of the control system has been investigated. When the parameter values are known, the controller in combination with the observer works adequately. The parameter estimator, however, yields incorrect parameters, which are caused by a coupling between two parameters. This causes a deviation of the estimated states from the process states. The simulation results suggest that a priori knowledge of the parameters is required for application of the model for control and state estimation.

Notes: Abstract In ethanol fermentations inhibition of the microorganism by ethanol limits the amount of substrate in the feed that can be converted. In a process high feed concentrations are desirable to minimize the flows. Such high feed concentrations can be realized in integrated processes in which ethanol is recovered from the fermentation broth as it is formed. In this study ethanol recovery by pervaporation was coupled to glucose fermentations by baker's yeast. Pervaporation was carried out with commercial silicone based hollow-fibre membrane modules with relatively high fluxes. Three different types of process configurations with pervaporation were investigated. Two of these configurations also included cell retention by microfiltration, in order to optimize the productivity. In the systems with pervaporation a feed containing 360 kg/m3 glucose could be converted almost completely. This feed concentration is a factor three higher than in a process without ethanol recovery. The productivity was 14 kg/m3 h in a system with pervaporation only, and could be increased to 43 kg/m3 h in the system with all recycle by microfiltration. The kinetic data suggest that accumulation of inhibitory compounds occurs in the integrated system. The integrated process was relatively easy in operation.

Notes: Abstract Four systems comprising of an ethanol fermentation integrated with microfiltration and/or pervaporation, and a conventional continuous culture, were compared with respect to the performance of the fermentation and economics. The processes are compared on the basis of the same kinetic model. It is found that cell retention by microfiltration leads to lower production costs, compared to a conventional continuous culture. Pervaporation becomes profitable at a high selectivity of ethanol/water separation and low membrane prices.

Notes: Abstract Membrane recycle fermentors are used successfully on laboratory scale to increase the efficiency of fermentation processes. The design of a process on larger scale however is obstructed by the lack of relevant data in literature. Compared to a stand-alone fermentor a membrane recycle fermentor presents new features which must be considered in the design. These features include the use of high density cultures, the additional volume in the membrane section and the circulation of the broth. In this theoretical study these aspects are analyzed with the characteristic time concept, in case of an ethanol fermentation integrated with microfiltration. The analysis shows that depending on the reactor configuration used concentration gradients can be expected. These gradients may decrease the efficiency of the fermentation, or can be advantageous, for example by letting the substrate conversion approach completion in the membrane section.