Christian Trägårdh

Physiological responses to mixing in large scale bioreactors

Escherichia coli fed-batch cultivations at 22 m3 scale were compared to corresponding laboratory scale processes and cultivations using a scale-down reactor furnished with a high-glucose concentration zone to mimic the conditions in a feed zone of the large bioreactor. Formate accumulated in the large reactor, indicating the existence of oxygen limitation zones. It is suggested that the reduced biomass yield at large scale partly is due to repeated production/re-assimilation of acetate from overflow metabolism and mixed acid fermentation products due to local moving zones with oxygen limitation. The conditions that generated mixed-acid fermentation in the scale-down reactor also induced a number of stress responses, monitored by analysis of mRNA of selected stress induced genes. The stress responses were relaxed when the cells returned to the substrate limited and oxygen sufficient compartment of the reactor. Corresponding analysis in the large reactor showed that the concentration of mRNA of four stress induced genes was lowest at the sampling port most distant from the feed zone. It is assumed that repeated induction/relaxation of stress responses in a large bioreactor may contribute to altered physiological properties of the cells grown in large-scale bioreactor. Flow cytometric analysis revealed reduced damage with respect to cytoplasmic membrane potential and integrity in cells grown in the dynamic environments of the large scale reactor and the scale-down reactor.

Substrate gradients in bioreactors: origin and consequences

Gradients of glucose in time and space are shown in a 30 m3 cultivation of Saccharomyces cerevisiae grown in minimal medium to a cell density of 20 gl−1. The fed-batch concept was used with glucose as the limiting component which was fed continuously to the process. As the mean glucose concentration declined throughout the process, the level of glucose was at all times different in three sampling ports (bottom/middle/top) of the reactor. These gradients were furthermore shown to depend on the feed position. This means that if the feed was supplied in the relatively stagnant mixing zone above the top impeller, the gradients were more pronounced than by feed in the well mixed bottom impeller zone. A rapid sampling system was constructed, and continuous glucose samples of every 0.15 s were analysed from a point of the reactor. Fifty samples were collected with this system, but the amount and frequency is possible to change. The results of these series show a variance of the glucose concentration where at one stage, a peak appeared of a relative difference in concentration of 40 mgl−1. The pattern of these rapid glucose fluctuations was shown to depend on the turbulence level at the location of the feed. It was shown, that the fluctuations were more pronounced when the feed was localised in a relatively stagnant area than in the well-mixed impeller area, where the deviation from the mean was negligible. The fluid flow, in the impeller (gassed and ungassed) and bulk area (ungassed) of the reactor, was characterised by turbulence measurements using thermal anemometry. These types of areas resembles well the different areas of sampling as mentioned above. The turbulent frequencies in these areas were in the range of 10−1 to 104 Hz with the highest amplitudes at low frequencies. The spectra depicts a uniform time scale for all zones, especially at the low frequencies. The dominance of low frequency, high amplitude flow variations and the observed short-time oscillations in substrate concentration support the hypothesis of substrate transport over fairly long distances without substantial mixing both in the impeller, but especially, in the bulk zone of the reactor.Simulations with an integrated CFD and biokinetic model were performed. The predictions of the glucose gradients of this model were compared to measurements.

Determining the zeta-potential of ceramic microfiltration membranes using the electroviscous effect

Abstract The possibility of measuring the zeta-potentials of porous membranes using the electroviscous effect was investigated. The zeta-potential of Membralox® ceramic microfiltration membranes was determined both with the newly developed electroviscous technique and by streaming potential measurements. It was found that the electroviscous technique provided a simple means of obtaining accurate values of zeta-potential, especially for higher zeta-potentials. The streaming potential measurements were found to be more suitable for the determination of the iso-electric point, i.e. the pH at which the zeta-potential is zero. The iso-electric points of new α-alumina, zirconia, and titania membranes were found to be 8.5, 8.0, and 6.3, respectively. Upon using the membranes and cleaning them with a detergent, the iso-electric point of the α-alumina membrane decreased to 6.5, and that of the zirconia membrane decreased to 5.2, while the iso-electric point of the titania membrane stayed virtually constant. Cleaning these membranes with a strong acid or base could not reverse the observed decreases in iso-electric point.

The influence of the membrane zeta potential on the critical flux for crossflow microfiltration of particle suspensions

Abstract It was investigated how the value of zeta potential of microfiltration membranes influenced the critical flux while filtering a suspension of silica particles. Ceramic membranes of three different materials were used and measurements were performed at two different values of pH for each material corresponding to two different values of zeta potential for each material. It was found that neither the zeta potential of the membrane nor the zeta potential of the particles influenced the observed critical flux. The critical flux increased linearly with the wall shear stress and decreased with increasing particle concentration. Expressions were obtained for the critical flux based on two different particle transport mechanisms: particle rolling (torque-balance model) and shear-induced diffusion. It was found that neither of the expressions could explain experimental critical fluxes adequately.

Production of W/O/W emulsions and S/O/W pectin microcapsules by microchannel emulsification

Abstract Water in oil in water (W/O/W) emulsions were produced by microchannel (MC) emulsification using water in oil (W/O) emulsions prepared by homogenization as feed emulsions. Polyoxyethylene sorbitan monolaurate (Tween 20) was used in the external water phase to stabilize oil droplets containing water droplets. Sorbitan monolaurate (Span 20), sorbitan monooleate (Span 80) and tetraglycerol polyricinoleate (TGPR) were tested as surfactants to stabilize the feed W/O emulsions. W/O/W emulsions were produced when using oleic acid or triolein as the oil phase and TGPR as the surfactant, while W/O/W emulsions could not be produced using Span 20 or Span 80 due to large water clusters and the low stability of the W/O emulsions. Using oleic acid as an oil phase, monodisperse W/O/W emulsions were obtained, while polydisperse W/O/W emulsions were produced using triolein as the oil phase, probably owing to the low production rate of the emulsion and fluctuation of the production rate. The concentration of TGPR affected the stability of the internal water droplets and oil droplets containing the water droplets. At a high TGPR concentration, the internal water was stable, however, oil droplets containing water droplets had a tendency to coalesce. W/O/W emulsions that contained pectin solution as an internal water phase were also produced using oleic acid and TGPR as the oil phase and surfactant in the oil phase, respectively. Solid in oil in water (S/O/W) pectin microcapsules were obtained by gelation of the internal water phase using a calcium solution containing Tween 20 as an external water phase.

Particle transport in crossflow microfiltration – II. Effects of particle–particle interactions

Abstract A model previously developed for the calculation of limiting fluxes for crossflow microfiltration of non–interacting particles was extended to include the effect of physico-chemical particle–particle interactions. It was shown theoretically that the effect of particle–particle interactions on the microfiltration flux can be described by a diffusion-type equation with an effective interactioninduced diffusion coefficient. The microfiltration flux for a general situation, where particle transport is caused by convection, Brownian diffusion, shear-induced diffusion, and particle–particle interactions, was then calculated by adding the diffusion coefficients, and solving the governing convective-diffusion equation numerically. Results of these calculations agreed very well with experimental fluxes measured during crossflow microfiltration of model silica particle suspensions. The influence of wall shear stress, membrane length, particle size, and particle concentration on permeate flux was very well predicted. However, the effect of particle surface potential was quantitatively underpredicted. Shear-induced diffusion seems to be the main transport mechanism governing the flux in the microfiltration of suspensions of micron-sized particles, but charge effects can increase fluxes considerably.

Time-Resolved Shear Viscosity of Wheat Flour Doughs—Effect of Mixing, Shear Rate, and Resting on the Viscosity of Doughs of Different Flours

ABSTRACT The shear viscosity of three doughs of different wheat cultivars mixed to a farinograph level of 500 BU was measured at low shear rates as a function of the shear deformation using a cone-and-plate viscometer. Cyanoacrylate adhesive was used to attach the dough samples to the instrument surfaces to eliminate wall slip. Flours used were Dragon, Kosack, and a fodder wheat. A distinct difference was observed between the viscosities of the different flour cultivars. The strongest dough (Dragon), with the highest protein content and a good resistance in the farinograph, had the highest maximum viscosity. The doughs showed distinct strain hardening, more pronounced for the strong doughs. Maximum viscosity was obtained at a strain of ≈4, almost independent of the shear rate, but at higher values for stronger doughs (5 for Dragon, 4 for Kosack, and 3.5 for fodder wheat). The maximum was most pronounced for well-mixed doughs after resting. The viscosity and its variation with strain may be used as a measu...

Mixing in industrial Rushton turbine-agitated reactors under aerated conditions

A new method of estimating the mixing time in industrial Rushton turbine-agitated reactors under aerated conditions is presented—called the analogy mixing-time model. It is based on the hypothesis of the analogy at the same turbine agitation speed between the mixing time in single liquid phase and that in two gas–liquid phases. This assumption is validated for production tanks equipped with two, three and four Rushton impellers. Furthermore, it is demonstrated that the internal geometry of the reactor, the number of impellers, the distance between impellers, etc., and the degree of homogeneity do not affect the model. Only the location at which the pulse is injected is found to influence the sensitivity coefficient . This result thus makes the analogy mixing-time model very useful in determining agitation parameters in production stirred reactor. The physical mechanisms behind the model are clarified with general gas–liquid flow patterns described in the literature. The study is completed by the description of the formation of concentration gradients in one of the reactors studied (12 m3 reactor). The effects of aeration on gradient formation are clarified with both bulk and impeller region injections. The results obtained are used for the design of a scaled-down reactor that mimics the environment created in the tank. (Less)

Computer simulations of mass transfer in the concentration boundary layer over ultrafiltration membranes

Abstract The phenomenon studied in this article is high-Schmidt-number mass transfer in the concentration polarization boundary layer during ultrafiltration (UF). The equations governing the transport of momentum, mass and concentration of a chemical species are solved throughout the boundary layer by a simulation program based on a control volume formulation. This formulation has the important quality that concentration-dependent physical properties can be calculated throughout the concentration boundary layer at spatially resolved points. To obtain a closeable set of momentum equations the turbulent transport of momentum is calculated using a slightly modified version of a low-Reynolds-number k -ϵ turbulence model formulated by Chien [1980]. The turbulent momentum transport is related to the turbulent transport of species concentration by a turbulent Schmidt number expression. The ability of the program to predict high-Schmidt-number mass transfer to a surface with suction is verified by comparison with experimental results. Thereafter the calculation procedure is used to predict mass transfer in UF, both for steady state and for a developing concentration polarization boundary layer. The simulation results are also compared with results obtained with the film theory equation, commonly used in membrane technology.

Particle transport in crossflow microfiltration – I. Effects of hydrodynamics and diffusion

The limiting flux in the crossflow microfiltration of particle suspensions was calculated by numerically solving the convective-diffusion equation governing concentration polarisation. It was assumed that a cake layer is formed at the membrane surface, and that particles are transported towards the membrane by convection, and transported away from the membrane by Brownian diffusion or shear-induced diffusion, or a combination of both these mechanisms. The numerical results for Brownian diffusion and for shear-induced diffusion could be summarised by approximate equations which related the predicted flux to parameters such as wall shear stress, bulk concentration, and membrane length. Both the approximate equations and the exact numerical results compared well with fluxes measured in crossflow microfiltration of non-interacting spherical silica particles. It was shown theoretically that turbulent diffusion contributes only marginally to particle transport in the concentration polarisation layer. Microfiltration fluxes for turbulent flow conditions were therefore predicted accurately by the numerical calculations, although turbulent diffusion was neglected.