Process Design Aspects for Small-Scale Fermentation Systems

Filamentous or biopolymer producing microorganisms are currently gaining importance due to their huge potential regarding worldwide sustainability issues. An economic industrial screening of these organisms is essential for determining the best producing microbial strains or fermentation conditions. It is known that filaments and biopolymers massively influence viscosity of the fermentation broth, thereby affecting the oxygen supply for microorganisms. In literature, the influence of viscosity is well documented for the oxygen transfer in stirred tank fermentations, however, little is known about the influence of viscosity on gas/liquid oxygen transfer in shake flask cultures. Even the effective shear rate in shake flasks was never systematically investigated, although it influences the apparent viscosity, mixing as well as mass and heat transfer. Since unknown oxygen transfer rates and unknown effective shear rates pose the risk of screening and producing under unfavorable conditions, this work addresses this lack of knowledge.Four objectives were pursued within this study. First, measurements and numerical simulations of the oxygen transfer in liquid films adhering on shake flask walls were conducted as a function of viscosity and film thickness. Thereby, the suitability of the widely applied film theory of Higbie was studied. 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 counter-intuitively increases in shake flasks when viscosity is increased from 1 mPa·s to 10 mPa·s, leading to an improved oxygen supply for microorganisms. Additionally, the OTRmax at viscosities of up to 80 mPa·s is not significantly lower than the OTRmax at waterlike viscosities. This is contrary to stirred tanks, where the oxygen supply is steadily reduced to only 5% at 80 mPa·s. Second, a first shear rate correlation for shake flasks – valid for a wide range of pseudo-plastic flow behaviors, shake flask sizes and operating conditions – was developed as a function of viscosity on the basis of Buckingham’s π-theorem and experimental data. It was found that effective shear rates in shake flasks commonly cover a range from 20 1/s to 2000 1/s. The precise applicability of the developed shear rate correlation was demonstrated for three different shake flask fermentations. Depending on the broth’s flow behavior, the effective shear rate in shake flasks is at least 1.55 times higher than that in stirred tank reactors operated at the same volumetric power input, leading to a potentially 50% lower apparent viscosity in shake flasks. Third, a novel concept of microtiter plates (MTP) in 96-well, 48-well and 24-well format was developed preventing spill-out of the rotating liquid even at high shaking frequencies and high filling volumes. In spite of the fact that high filling volumes per well are often desirable for offline analytics during screening procedures, as yet, low filling volumes per well have to be adjusted to prevent spill-out and to overcome insufficient oxygen transfer in conventional MTPs. In prototypes of the newly developed MTP, preliminary measurements of the OTRmax were conducted using a sulfite system. With respect to the OTRmax, the advantages of the novel MTP could clearly be displayed for ranges of filling volume and shaking frequency where conventional MTPs spill out: Whereas extreme shaking conditions are needed for the novel 96-well type to reach a higher OTRmax compared to the conventional one, the novel 48-well and 24-well formats are already advantageous at moderate shaking conditions. This is due to the influence of the surface tension which becomes less dominant the larger the well diameter is. The biocompatibility of the novel MTP type was proven by Escherichia coli fermentations. Due to higher achievable oxygen transfer rates and, thus, faster C-source consumption compared to cultures in the conventional MTP, the new MTP even showed the potential of shortening the fermentation time. Hence, a suitable MTP concept was found for applications in high-throughput screenings where high sample volumes are required.Fourth, a scale-down was conducted requested by an industrial collaboration partner. To achieve time and cost efficient high-throughput for strain screening, an established shake flask protocol was scaled down into MTP. An approach based on an oxygen-consuming sulfite system was applied to ensure equal OTRmax-values in MTPs and shake flasks. Obtained sulfite datasets were used to identify operating conditions leading to the same oxygen supply for the model organism Trichoderma reesei in shake flasks and 24-well MTPs. For 24-well MTPs, the shake flask OTRmax of 20 mmol/L/h of the industrial protocol was obtained under the following optimal operating conditions: 1 mL filling volume per well, 200 rpm shaking frequency and 50 mm shaking diameter. With the identified operating conditions almost identical oxygen transfer rates and product concentrations were measured in shake flasks and 24-well MTP cultures as a function of fermentation time. The proposed sulfite approach is a fast and accurate means to scale-down established screening procedures into MTPs to achieve high-throughput. The obtained insights into oxygen transfer and effective shear rates in viscous systems are valuable for explaining existing deviations in screening and production results. Ultimately, by means of consistent scale-up and scale-down procedures, economic bioprocess development is facilitated with the results of this study