Applying Computational Fluid Dynamics Technology in Bioprocesses-Part 2 - Computational fluid dynamics can resolve performance problems. - BioPharm International


Applying Computational Fluid Dynamics Technology in Bioprocesses-Part 2
Computational fluid dynamics can resolve performance problems.

BioPharm International
Volume 23, Issue 5


It has been observed that the biomass yield on the carbon and energy source is decreased by the scale-up of aerobic processes. For example, in an E. coli-based recombinant protein process, the biomass yield of glucose and the maximum cell density reached in the process dropped by about 20% when scaling up from 3 to 9,000 L.55 In large-scale fed-batch processes, poor mixing may cause a large gradient of the concentration of the limiting substrate. Streptomyces fradiae, has been shown to have a change of 55-fold in the rates of oxygen and nutrient consumption as it circulates through different micro-environments of varying concentrations of oxygen and nutrient.56 Two possible scenarios may come up—the formation of moving oxygen-limiting zones when the cell density is high enough, caused by the fast use of sugar by cells, and the occurrence of overflow metabolism responded by the many microorganisms being exposed to glucose concentration above a critical level—Saccaromyces cerevisiae producing ethanol and E. coli-producing acetate.25,57 Furthermore, the occurrence of oxygen-limitation may result in anaerobic reactions and concomitant stress responses.58

Aeration is an essential requirement for aerobic cell lines. However, DO gradients are likely to occur when the characteristic time for oxygen uptake is shorter than the mixing time. In aerobic fermentation systems, the rate of oxygen transfer to the cells usually is the limiting factor. A key factor that influences oxygen transfer is bubble size distribution, which depend on turbulent characteristics in the reactor.59 The bubble sizes dictate the available interfacial area for gas–liquid mass transfer. The scale-up and design of bioreactors must meet oxygen transfer requirements while maintaining low shear rates and a controlled flow pattern. Therefore, understanding the hydrodynamic environment is critical for fermentation processes.60

One of the challenges for fermentation process scale-up is to prevent concentration gradients (from the feed zone to the interior of the tank) resulting in further limitations and unwanted byproduct formation or inhibition of product rate. The following examples indicate that CFD simulations can be applied for the optimization of fed batch processes to avoid oxygen limitations, overflow metabolism, or catabolite regulation.

Large-Scale Tanks

Model simulations for large-scale tanks with multiple Rushton and axial flow impellers of different combinations shows that feeding the concentrated sugar solution into the impeller region and axial flow impellers leads to the best equidistribution of substrate.61 Bacillus subtilis is used as a model for an oxygen sensitive culture for the characterization of the effects of inhomogeneities on the intensity of oxygen transfer gas liquid in stirred-tank reactor. CFD predictions for DO concentration and the production rate of butanediol show qualitative agreement with experimental observations.61

Gradients of glucose in time and space are found in a 30 m3 fed-batch cultivation of S. cerevisiae grown in minimal medium to a cell density of 20 g/1, which were shown to depend on the feed position. Simulations with an integrated CFD and biokinetic model are performed to investigate the glucose gradients of this model.62 The simulations show a glucose gradient formed according to a change in the point of feeding and caused by a different cell concentration. The glucose concentration is high near the feed inlet and declines further away. The effect of the feed position appears as a generally higher mean concentration for the top feed compared with the bottom.


It is not rare to observe abnormal performances such as lower cell viabilities, higher death rates, and a drop of titer when cell culture processes are scaled up from laboratory-scale to pilot or product scale. The potentially lethal events in a mechanically agitated and sparged bioreactor have been identified.63–66 They are bubble formation and detachment at the sparger, bubble interaction with the impeller in the impeller zones, bubbles rising along their path surfacing to the interface, and bubbles bursting at the media–air interface. Sparging gas into a tank generates strong turbulent flow and high shearing force at the sparger.67 Without surfactant additives such as Pluronic F68, cells tend to attach to bubbles and thus experience shearing forces in the thin film surrounding rising bubbles and in the wake behind bubbles.68–71

A complex dynamic takes place in the impeller zones when stable air-filled cavities are formed. Centrifugal force competes with buoyancy force in the impeller zones, of which the coupling interactions determine the bubble residence time, bubble breakup frequency, and bubble size distribution.52,72 The existence of the medium–air interface has a large effect on cell damage.73–80 If the air–medium interface at the top of the tank is eliminated, no cell damage is observed at impeller speeds up to 600 rpm for hydridoma cells.81 When a bubble approaches the air–liquid interface and subsequently bursts as a result of film draining, the shear stress in the resultant rimming flow is estimated at up to 200 Pa.75 On average, about 1,000 cells are killed per 2 mm of bubble rupture.79 Computer simulations of bubble rupture processes indicate that the rupture of small air bubbles in pure water generates an energy dissipation rate three to four orders of magnitude higher than that typically created in bioreactors solely because of agitation.82–84 The energy dissipation rate increases rapidly with a decrease in bubble size.

CFD can provide information on the parameters most closely linked to cell damage, such as shear stress, power consumption, Kolmogrov eddy length scale distribution, turbulence characteristics including Reynolds stress, kinetic energy, and energy dissipation rate.85,86 CFD models also can predict process conditions that will cause vortices to occur and also the product vortex strength, compute cell residence time, and provide bubble size distribution, etc. With these kinds of information, one can easily trouble-shoot process performance issues. CFD simulation results show that shear force on the impeller surface, Reynolds stress, and turbulence energy dissipation rate in the impeller zones are unlikely to cause cell damage under typical cell culture process conditions. On the other hand, the interaction between the impeller and bubbles, which leads to bubble breakup in the impeller zones, and bubble bursting at the air–media interface are more likely to be responsible for observed cell damage.87,88


A detailed review has been provided of how computational fluid dynamics (CFD) can be used to analyze bioreactor and fermentater unit operations, including flow characterization, mixing, resuspension, scale-up, and cell damage. CFD technology is a useful approach that provides detailed information that helps process development scientists understand cell culture and fermentation process performance problems encountered during technology transfer. However, some of the studied cases also indicate the inadequacy of the modeling efforts because of the complexity of biological systems and inaccurate physical models. Understanding a process physically and biologically is key to the success of the CFD modeling and simulations with regard to biopharmaceutical processes.

Zhiwu (David) Fang is the founder and principal consultant of Systems Quality-by-Design, Inc., Newbury Park, CA, 626.716.2016,

blog comments powered by Disqus



Lundbeck CEO Resigns Due to Code of Conduct Breach
November 24, 2014
GPhA Issues Statement on Generic Drug Costs
November 20, 2014
Amgen Opens Single-Use Manufacturing Plant in Singapore
November 20, 2014
Manufacturing Issues Crucial to Combating Ebola
November 20, 2014
FDA Requests Comments on Generic Drug Submission Criteria
November 20, 2014
Author Guidelines
Source: BioPharm International,
Click here