Liquid films on shake flask walls explain increasing maximum oxygen transfer capacities with elevating viscosity.

In biotechnological screening and production, oxygen supply is a crucial parameter. Even though oxygen transfer is well documented for viscous cultivations in stirred tanks, little is known about the gas/liquid oxygen transfer in shake flask cultures that become increasingly viscous during cultivation. Especially the oxygen transfer into the liquid film, adhering on the shake flask wall, has not yet been described for such cultivations. In this study, the oxygen transfer of chemical and microbial model experiments was measured and the suitability of the widely applied film theory of Higbie was studied. With numerical simulations of Fick's law of diffusion, it was demonstrated that Higbie's film theory does not apply for cultivations which occur at viscosities up to 10 mPa s. For the first time, it was experimentally shown that the maximum oxygen transfer capacity OTRmax increases in shake flasks when viscosity is increased from 1 to 10 mPa s, leading to an improved oxygen supply for microorganisms. Additionally, the OTRmax does not significantly undermatch the OTRmax at waterlike viscosities, even at elevated viscosities of up to 80 mPa s. In this range, a shake flask is a somehow self-regulating system with respect to oxygen supply. This is in contrary to stirred tanks, where the oxygen supply is steadily reduced to only 5% at 80 mPa s. Since, the liquid film formation at shake flask walls inherently promotes the oxygen supply at moderate and at elevated viscosities, these results have significant implications for scale-up.

New method to determine the mass transfer resistance of sterile closures for shaken bioreactors.

In this paper a novel and easily applied method to measure the mass transfer resistance of the sterile closures (e.g. cotton plug) of shaken bioreactors is introduced. This method requires no investment in special equipment (e.g. an oxygen sensor) and can be performed with the materials usually available in typical laboratories. The method is based on the model of Henzler et al. (1986), which mechanistically describes mass transfer through the sterile closure of a shaken bioreactor based on diffusion coupled with Stefan convection. The concentration dependency of the multi-component diffusion coefficients is taken into account. The water loss from two equivalent shaken bioreactors equipped with sterile closures during several days of shaking is measured. One flask contains distilled water, the other a saturated salt solution. From the water evaporation rate in each of the two flasks, the new model presented calculates the relative humidity in the environment, the average diffusion coefficient of oxygen in the sterile closure (D(O2)), and the diffusion coefficient of carbon dioxide (D(CO2)) . The diffusion coefficient of carbon dioxide (D(CO2)) only depends on the density and material properties of the sterile closure and not on the gas concentrations and is, therefore, an ideal parameter for the characterization of the mass transfer resistance. This new method is validated experimentally by comparing the diffusion coefficient of oxygen (D(O2)) to a measurement by the classic dynamic method; and by comparing the calculated relative humidity in the environment to a humidity sensor measurement.

Hydromechanical stress in shake flasks: correlation for the maximum local energy dissipation rate.

Shake flasks are widely used in biotechnological process research. Bioprocesses for which hydromechanical stress may become the rate controlling parameter include those where oils are applied as carbon sources, biotransformation of compounds with low solubility in the aqueous phase, or processes employing animal, plant, or filamentous microorganisms. In this study, the maximum local energy dissipation rate as the measure for hydromechanical stress is characterized in shake flasks by measuring the maximum stable drop size. The theoretical basis for the method is that the maximum stable drop diameter in a coalescence inhibited liquid/liquid dispersion is only a function of the maximum local energy dissipation rate and not of the dispersing apparatus. The maximum local energy dissipation rate is obtained by comparing the drop diameters in shake flasks to those in a stirred tank reactor. At the same volumetric power consumption, the maximum energy dissipation rate in shake flasks is about 10 times lower than in stirred tank reactors explaining the common observation of considerable differences in the morphology of hydromechanically sensitive cells between these two reactor types. At the same volumetric power consumption, the maximum local energy dissipation rate in baffled and in unbaffled shake flasks is very similar. A correlation is presented to quantify the maximum local energy dissipation rate in shake flasks as a function of the operating conditions. Non-negligible drop viscosity may be considered by known literature correlations. Further, from dispersion experiments a critical Reynolds number of about 60,000 is proposed for turbulent flow in unbaffled shake flasks.