A continuous,multistage tower fermentor. I. Design and performance tests

The design of a continuous column fermentor with a multiple staging effect is described. The column is divided into four compartments by horizontal perforated plates and is provided with a central agitator shaft driving an impeller in each compartment. A tube at the center of each plate forms a liquid seal around the shaft and also acts as a “downcomer.” The fermentor is normally operated with counter-current flow of gas and medium. Fresh medium is added to the top stage and product is withdrawn from the bottom. The effect of plate and agitator design on fermentor performance was studied in terms of factor such as oxygen transfer rate, gas holdup, and interstage mixing. By proper choice of the design parameters, the fermentor was made to approximate a perfect four-stage cascade in terms of reactor performance. Preliminary experiments were performed with air-water systems, but a more realistic picture of fermentor performance was obtained in experience involving propagation of Escherichia coli. Data for business and substrate concentrations in each stage confirmed the staging effect of the apparatus. The fermentor operated in a stable manner for periods of more than two weeks.     

Modelling mass transfer and agitator performance in multiturbine fermentors

A methodology for mathematically analyzing agitator performance and mass transfer in large multiturbine production fermentors is presented. The application of this approach provides a method for determining axial dissolved oxygen profiles under conditions of known mass transfer rates as a function of agitation-aeration characteristics. A stagewise approach is used which divides the fermentor into a series of mixing cells. This allows for each turbine and mixing cell to be individually optimized. The model also permits the determination of the mass transfer coefficient for each turbine based upon limited dissolved oxygen data. The primary limitation of this approach rests in the limited data and correlations available for multiturbine systems. The structure of the modelling approach can serve as a basis for testing single turbine correlations and adapting them to multiturbine systems. The step-by-step details of the mathematical analysis are presented and interpreted. A series of computer simulations demonstrate the effect of typical fermentor operating variables on the axial dissolved oxygen profile. Further simulations demonstrate the effect of modifying agitator blade numbers on the dissolved oxygen profile and agitator power requirement.     

The influence of impeller type in pilot scale xanthan fermentations

The rheological complexity of Xanthan fermentations presents an interesting problem from a mixing viewpoint, because the phenomena of poor bulk blending and low oxygen mass transfer rates inherent in highly viscous fermentations (and their consequences) can be systematically investigated, even at the pilot plant scale. This study in a 150 L fermentor compares the physical and biological performance of four pairs of impellers: a standard Rushton turbine, a large diameter Rushton turbine, a Prochem Maxflo T, and a Scaba 6SRGT. Accurate in-fermentor power measurements, essential for the comparison of impellers in relation to operating costs are also reported. It is demonstrated that the agitator performance in Xanthan fermentations is very specific and the choice of which impeller to use in bioreactors to obtain enhanced performance is dependant on the applied criterion. None of the criterion favored the use of the standard Rushton turbine, therefore suggesting that there are strong grounds for retrofitting these impellers with either large diameter impellers of similar design or with novel agitators. In addition, fluid dynamic modeling of cavern formation has clearly highlighted the importance of a well mixed and oxygenated region for providing the capacity for high microbial oxygen uptake rates which govern Xanthan productivity and quality. Copyright 1998 John Wiley & Sons, Inc.     

翼形轴流桨在红霉素发酵中的应用研究

在20t工业发酵罐中,研究了翼形轴流桨搅拌对红霉素发酵过程的影响。重点考察了粘度,溶氧,效价等过程参数变化,以及搅拌功耗与发酵产量之间的关系。研究结果表明：（1）与传统的涡轮搅拌桨相比,翼形轴流桨搅拌其发酵过程参数（粘度,溶氧,效价等）随时间的变化曲线有明显的差异;（2）在相同的生产条件下,用翼形轴流桨代表涡轮桨可有效提高生产效率。     

Distributed parameter model of an airlift fermentor

A distributed parameter model for an airlift fermentor is presented. A riser represents the airlift fermentor, with plug flow in both gas and liquid phases, a well-mixed section that acts as gas separator, and a downcomer with plug flow. The set of equations proposed makes possible both the understanding and design of the system. Macroscopic balances shows a behavior that is very close to conventional continuous stirred tank fermentor from the viewpoint of biomass production. In addition, the model predicts concentration profiles of biomass, substrate and oxygen in the liquid, and oxygen in the gas phase. This allows estimation of optimal gas flow rate for sufficient oxygen transfer with minimum energy input.     

Comparisons between cellulase production by Aspergillus fumigatus in agitated vessels and in an air-lift fermentor

Aspergillus fumigatus was cultured in disc-turbine-agitated vessels and in an air-lift fermentor. In the agitated vessels the yield of cellulase was reduced when the agitation rate was increased, although extracellular protein levels rose. The enzyme complex itself was shown to be exceptionally stable under conditions similar to those in the agitated vessels, so probably shear damage to the mycelium had occurred, liberating intracellular contents. These appeared to contain an inhibitor that could be removed by fabricated inorganic protein absorbents, such as kieselguhr and alumina. However, the inhibitor was not likely to be protease, since only relatively low levels could be detected and its identity has not been established. The use of an air-lift fermentor avoided the shear effects due to use of the disc turbine agitator in the conventional fermentors, and yields of enzyme were then found to increase by about 20%, maximum yields being obtained at maximum K(L)a values.     

Two- and three-phase mixing in a concentric draft tube gas-lift fermentor

Two- and three-phase mixing studies were carried out in a 44-L concentric draft tube gas-lift fermentor. It was proposed to use the fermentor for the production of solvents using immobilized bacteria. Bubble size, gas holdup, liquid velocities, circulation, and mixing times were determined for various superficial gas velocities in distilled water, starch, carboxymethyl cellulose, and ethanol solutions. The observed trends for two phase mixing were similar to other studies but the results were found to be more sensitive to liquid properties. This was possibly due to the large value of downcomer to riser area used in this study. Mixing in three phases highlighted the difficulty in predicting the effect of adding solids to the gas-liquid system. Results showed that the gas-lift fermentor was ideally suited to dealing with three phases but more work is necessary before accurate models can be developed to account for the effect of solids.     

Liquid level control in an aerated continous-flow fermentor

An electronic liquid level control system was developed to maintain constant working volume in a continuously fed and aerated fermentor from which spent medium was continously withdrawn by microfiltration in an external recycle loop. The level control system was comprised of an admittance sensor in the fermentor and an external transmitter and controller, which together regulated the speed of a recycle pump and thus the rate of liquid withdrawal from the fermentor. During test bacterial culture the liquid level was maintained usually within +/- 1% and always within +/- 5%. The control system could be applied to various other types and scales of continuous-flow fermentors.     

Computational analysis of confined jet flow and mass transport in a blind tube.

A computational analysis of confined nonimpinging jet flow in a blind tube is performed as an initial investigation of the underlying fluid and mass transport mechanics of tracheal gas insufflation. A two-dimensional axisymmetric model of a laminar steady jet flow into a concentric blind-end tube is put forth and the governing continuity, momentum, and convection-diffusion equations are solved with a finite element code. The effects of the jet diameter based Reynolds number (Re(j)), the ratio of the jet-to-outer tube diameters (epsilon), and the Schmidt number (Sc) are evaluated with the determined velocity and contaminant concentration fields. The normalized penetration depth of the jet is found to increase linearly with increasing Re(j) for epsilon = O(0.1). For a given epsilon, a ring vortex that develops is observed to be displaced downstream and radially outward from the jet tip for increasing Re(j). The axial shear stress profile along the inside wall of the outer tube possesses regions of fixed shear stress in addition to a local minimum and maximum in the vicinity of the jet tip. Corresponding regions of axial shear stress gradients exist between the fixed shear stress regions and the local extrema. Contaminant concentration gradients develop across the ring vortex indicating the inward diffusion of contaminant into the jet flow. For fixed epsilon and Sc and Re(j) approximately 900, normalized contaminant flow rate is observed to be approximately twice that of simple diffusion. This model predicts modest net axial contaminant transport enhancement due to convection-diffusion interaction in the region of the ring vortex.     

Bubbling and foaming assisted clearing of mucin plugs in microfluidic Y-junctions

Microfluidic Y-junctions were used to study mechanical mechanisms involved in pig gastric mucin (PGM) plug removal from within one of two bifurcation branches with 2-phase air and liquid flow. Water control experiments showed moderate plug removal due to shear from vortex formation in the blockage branch and suggest a PGM yield stress of 35 Pa, as determined by computational fluid dynamics. Addition of hexadecyltrimethylammonium bromide (CTAB) surfactant improved clearing effectiveness due to bubbling in 1 mm diameter channels and foaming in 500 μm diameter channels. Plug removal mechanisms have been identified as vortex shear, bubble scouring, and then foam scouring as air flow rate is increased with constant liquid flow. The onset of bubbling and foaming is attributed to a flow regime transition from slug to slug-annular. Flow rates explored for 1 mm channels are typically experienced by bronchioles in generations 8 and 9 of lungs. Results have implications on treatment of cystic fibrosis and other lung diseases.     

A constant-depth laboratory model film fermentor

A laboratory model constant-depth film fermentor was developed. Film grew on the surface of polytetrafluoroethylene (PTFE) plugs and was limited to a predetermined depth by mechanically removing excess film. Six film-forming organisms were isolated from river water and used to assess the operating characteristics of the fermentor. Film accumulation was logarithmic, and a steady state was maintained. Electron micrographs show early film development. The fermentor enables film to be grown on any substratum and allows discrete, reproducible, and representative samples to be taken.     

Critical evaluation of sampling techniques for residual glucose determination in carbon-limited chemostat culture of Saccharomyces cerevisiae

In this paper, three sampling techniques for rapid quenching of cellular metabolism and subsequent separation of cells from fermentation broth are compared: (i) quick freezing of fermentation broth directly in liquid nitrogen; (ii) quenching metabolism by exposing the fermentation broth to stainless steel beads (4-mm diameter) in a filter syringe precooled to -18 degrees C; and (iii) withdrawal of the filtrate through a 0.45 microm filter attached to a syringe and a needle inserted directly into the fermentor. It was concluded that use of liquid nitrogen as a quenching method to rapidly arrest cellular metabolism, for quantitative analysis of extracellular glucose, is not a very reliable method and that the filter syringe steel beads work very well.     

不同搅拌器类型对齿毛菌发酵产漆酶的影响

【目的】白腐菌Cerrena unicolor Y-G07是一种不产孢子的丝状真菌,其漆酶合成与生长相偶联。应用于单细胞微生物发酵的搅拌器并不适用于白腐真菌。在此,综合考虑溶氧效果和剪切力作用,研究不同搅拌器类型对白腐菌Cerrena unicolor Y-G07发酵生产漆酶产量的影响。【方法】针对Y-G07菌株生长过程需氧量大,且对剪切力敏感的特性,订制5种不同类型的搅拌器(径向流、轴向流等),在通气已控制在设备最大量程及转速优化的前提下,研究Y-G07在发酵生产过程中使用不同类型搅拌器对菌丝生长形态、生长速度、溶氧情况、糖代谢和漆酶合成的影响。【结果】Cerrena unicolor Y-G07菌株对不同类型搅拌器产生的发酵液流态性质和剪切力敏感,表现在菌丝体的生长形态、细胞浓度差异较大,且生长周期改变,从而影响漆酶的合成。采用六折叶DT602搅拌器最有利于该菌株形成致密度合适的网状菌丝体,菌丝体细胞浓度高,断裂少、生长状况好,漆酶的单位产量可达690 U/m L,相对于普通使用的六叶平直叶搅拌器(448 U/m L)提高了54%。【结论】选择合适的搅拌器类型有利于好氧但对剪切力敏感的微生物发酵。     

Mechanism of uptake of liquid hydrocarbons by microorganisms

Growth rates of Candida tropicalis were studied in two different fermentors. One was the ordinary shaker flask containing both the aqueous culture medium and liquid hydrocarbons. The other was a specially designed rotating disk-type fermentor containing only the aqueous culture medium, into which vapors of n-paraffins from C₆ to C₁₈ were supplied continuously without forming the liquid hydrocarbon phase. The specific growth rates of Candida tropicalis in the rotating disk fermentor, under such conditions that supply of hydrocarbon vapor was sufficient, showed good agreement with those in the shaker flask. This seems to indicate that hydrocarbon uptake by Candida tropicals by direct contact with liquid hydrocarbon is negligible.     

Vortex Formation with a Snapping Shrimp Claw

Snapping shrimp use one oversized claw to generate a cavitating high speed water jet for hunting, defence and communication. This work is an experimental investigation about the jet generation. Snapping shrimp (Alpheus-bellulus) were investigated by using an enlarged transparent model reproducing the closure of the snapper claw. Flow inside the model was studied using both High-Speed Particle Image Velocimetry (HS-PIV) and flow visualization. During claw closure a channel-like cavity was formed between the plunger and the socket featuring a nozzle-type contour at the orifice. Closing the mechanism led to the formation of a leading vortex ring with a dimensionless formation number of approximate ΔT*≈4. This indicates that the claw might work at maximum efficiency, i.e. maximum vortex strength was achieved by a minimum of fluid volume ejected. The subsequent vortex cavitation with the formation of an axial reentrant jet is a reasonable explanation for the large penetration depth of the water jet. That snapping shrimp can reach with their claw-induced flow. Within such a cavitation process, an axial reentrant jet is generated in the hollow cylindrical core of the cavitated vortex that pushes the front further downstream and whose length can exceed the initial jet penetration depth by several times.     

Dissolved oxygen measurements in pilot- and production-scale novobiocin fermentations

Dissolved oxygen measurements were made in pilot (20 and 250 l.) and production scale (15,000 l.) novobiocin fermentations. Bulk mixing was found to be incomplete in pilot tanks with turbine impellers of D/T = 0.40 (where D is impeller diameter, and T is tank diameter) but appeared homogeneous with impellers of D/T = 0.69. In the former case, the respiration rate was presumably limited by insufficient oxygen supply in areas of poor bulk mixing, whereas, in the latter case, the major resistance was between the bulk of the liquid and the cell (intraclump resistance). Higher agitator speeds decreased the gas–liquid resistance proportionally more than they reduced the liquid-cell resistance. In production fermentors, dissolved oxygen measurements indicated that bulk mixing was complete with each of the three impeller sizes tested (D/T = 0.28, 0.33, and 0.43), but that the respiration rate was again limited, mainly by a resistance between the bulk of the liquid and the cell.     

Protease production by fermentation of fish solubles from salmon canning processes

Production of protease by fermentation, using Sorangium 495, of a substrate based on condensed fish solubles is demonstrated. The effects of carbohydrate addition, pH, fish solubles concentration, scale-up, agitation, and air flow rate on protease yields are described. While the fish solubles medium alone could give rise to measurable yields of protease, these were, at worst, doubled when 1% glucose was added to the medium. pH 7 was optimal for protease yield. Although the concentration of fish solubles in the basic medium showed no significant effect on cell yield, maximum protease yield was observed at a protein concentration equivalent to 3.85 mg/mL of bovine serum albumin. Protease production rates decreased as medium protein fermentor showed no significant effect on maximum protease yields. The effects of agitator speed and air flow rate on protease yield suggested that the rate of O2 transfer from air to medium could limit the rate of protease production. It was also noted that protease production is not growth associated.     

Miniaturized device for agitating a high-density yeast suspension that is suitable for in vivo nuclear magnetic resonance applications

In vivo nuclear magnetic resonance (NMR) monitoring requires a high-density cell suspension, where cell precipitation should be avoided. We have designed a miniaturized cell agitator that fits entirely into an 8-mm NMR probe but that, being mounted into the instrument, is situated outside of the sensitive area. The device consists of two glass tubes connected in a way that, when gas flow is blown through them, creates influx of cell suspension into the device that returns through apertures. This flow creates continuous circular vortex of the cell suspension in the whole sample volume, whereas there are no moving mechanical parts or gas bubbles crossing the instrument’s sensitive area. The gas flow controls conditions of the cell suspension and removes volatile waste metabolites.     

Integration of microbial kinetics and fluid dynamics toward model‐driven scale‐up of industrial bioprocesses

Scale‐up of bioprocesses is hampered by open questions, mostly related to poor mixing and mass transfer limitations. Concentration gradients of substrate, carbon dioxide, and oxygen in time and space, especially in large‐scale high‐cell density fed‐batch processes, are likely induced as the mixing time of the fermentor is usually longer than the relevant cellular reaction time. Cells in the fermentor are therefore repeatedly exposed to dynamic environments or perturbations. As a consequence, the heterogeneity in industrial practices often decreases either yield, titer, or productivity, or combinations thereof and increases by‐product formation as compared to well‐mixed small‐scale bioreactors, which is summarized as scale‐up effects. Identification of response mechanisms of the microorganism to various external perturbations is of great importance for pinpointing metabolic bottlenecks and targets for metabolic engineering. In this review, pulse response experimentation is proposed as an ideal way of obtaining kinetic information in combination with scale‐down approaches for in‐depth understanding of dynamic response mechanisms. As an emerging tool, computational fluid dynamics is able to draw a holistic picture of the fluid flow and concentration fields in the fermentor and finds its use in the optimization of fermentor design and process strategy. In the future, directed strain improvement and fermentor redesign are expected to largely depend on models, in which both microbial kinetics and fluid dynamics are thoroughly integrated.     

Experimental evaluation of liquid film resistance in oxygen transport to microbial cells

A membrane probe was used to monitor the dissolved oxygen concentrations in continuous cultures of Candida utilis and Micrococcus roseus growing at low dissolved oxygen concentrations and various agitation levels. For the yeast fermentations, increasing the agitation level within the range of 0.1 to 0.3 w per liter lowered steady-state dissolved oxygen concentrations in the fermentor. The steady-state dissolved oxygen concentration in the fermentor was not influenced by the agitation level within the range of 0.3 to 1.8 w per liter. With M. roseus, no effect of agitation on steady-state dissolved oxygen concentrations in the fermentor was observed within the range of 0.1 to 1.8 w per liter. It was concluded that, under the conditions used, a measurable transfer barrier from the liquid to the yeast cells existed at agitation levels below 0.3 w per liter and that this barrier did not exist at agitation levels above 0.3 w per liter. The transfer barrier from the liquid to the yeast surface could be represented by a stagnant film of liquid 0.6 × 10^-4 cm thick surrounding the cell at an agitation level of 0.10 w per liter. This film represented an oxygen concentration drop of 1.3 × 10^-7 M from the bulk of the medium to the cells under the experimental conditions.