In aerobic fermentation and cell‑culture processes, oxygen constitutes one of the most critical nutrients. Due to the extremely low solubility of oxygen in water, gas‑liquid mass transfer frequently becomes the bottleneck restricting cell growth and product formation. Oxygen cannot be simply “pumped into” the broth. Oxygen from air must go through a series of sequential steps: bubble generation, bubble breakup, gas‑liquid interfacial exchange, oxygen diffusion across liquid films, and finally uptake into cells. Limitations at any of these steps will result in insufficient oxygen availability for cells. This article systematically sorts out these key contributing factors to provide references for bioreactor design and process‑scale optimization.
Bubble Formation
A ring sparger is installed directly beneath the bottom impeller. Multiple small orifices are evenly distributed on the lower side or both sides of the ring pipe. When high‑pressure sterile air is fed into the ring, gas jets emerge from the orifices to form an initial bubble swarm, which is subsequently shattered by the impeller and dispersed throughout the fermentation broth. Gas dispersion inside the reactor mainly occurs in the vicinity of the impeller. Since bubbles generated at the sparger are immediately drawn into the impeller zone, the degree of gas dispersion within the reactor is largely independent of sparger geometry. Increased agitation speed enhances gas recirculation inside the vessel. Coalescence of small bubbles into large ones impairs oxygen transfer, as both total interfacial area and gas hold‑up decline. The coalescence behaviour of liquids depends on medium composition. Compared with pure water, the presence of salts and ions suppresses coalescence and favours oxygen transfer. Adding ions to water in sparged vessels substantially reduces average bubble size and elevates gas hold‑up; consequently, the specific interfacial area a in saline media can be up to 10‑fold higher than that in salt‑free water.
Bubble Morphology
Gas‑liquid mass‑transfer efficiency is strongly governed by the properties of bubbles dispersed within the liquid medium. Large‑scale aerobic cultivation is mostly carried out in stirred‑tank bioreactors, where oxygen is supplied by sparging gas beneath the impeller. The impeller functions to distribute bubbles across the entire reactor volume. In small laboratory‑scale bioreactors, nearly all liquid resides close to the impeller, so bubbles in these systems frequently interact with turbulent liquid flow. In contrast, after initial dispersion by the impeller in most industrial bioreactors, bubbles spend most of their time freely rising with little disturbance. Owing to scale‑up effects, fluids far from the impeller in large bioreactors receive insufficient energy for bubble breakup. Virtually all large‑scale commercial stirred‑tank reactors operate predominantly under conditions of freely rising bubbles.
Bubble size represents the most important bubble characteristic in bioreactors. For a given gas volume, dispersing gas into numerous small bubbles rather than a few large ones yields a much larger interfacial area. Therefore, a major objective in bioreactor design is to achieve high‑level gas dispersion. Nevertheless, small bubbles offer additional benefits: they exhibit lower rise velocities and longer liquid residence times, providing extended periods for oxygen mass transfer. A common misconception is that enhanced oxygen transfer is achieved merely by generating ever‑smaller bubbles. In reality, infinitely fine bubbles do not guarantee improved mass‑transfer performance. Once bubble size falls below a certain threshold, stable liquid films form at bubble surfaces and impede gas exchange. Accordingly, industrial fermenter design does not pursue minimal bubble size; instead, it targets a balanced trade‑off among bubble dimension, residence time and interfacial renewal rate.
Liquid Viscosity
Liquid properties exert pronounced impacts on bubble size, gas hold‑up and volumetric mass‑transfer coefficient kₗa. Increasing viscosity thickens the fluid boundary layer surrounding bubbles and raises resistance to oxygen transfer. High viscosity also suppresses turbulence, modifies the geometry and structure of impeller gassed cavities, and degrades gas‑dispersion efficiency. The kₗa value generally decreases as liquid viscosity rises. As illustrated in Figure 1, under identical bioreactor power input, kₗa values in carboxymethyl‑cellulose solutions are markedly lower than those measured in water or salt solutions.
Agitation Speed and Gas Flow Rate
Under typical bioreactor operating conditions, elevating agitation speed and gas flow rate improves kₗa. Typical kₗa datasets for Newtonian fluids of different viscosities are plotted in the accompanying figures. Figure 2A demonstrates the strong dependence of kₗa on agitation rate: in the studied system, doubling agitation speed increased average kₗa by a factor of 4.6. By comparison, increasing gas flow rate is a less effective strategy for boosting kₗa; for instance, doubling gas flow only increased kₗa by approximately 20%.
In most systems, there is limited practical scope to raise kₗa via higher gas flow. Unless the impeller can disperse all gas impinging on it, impeller flooding occurs at excessive gas flow rates. Large volumes of gas accumulate around the impeller to form a substantial “gas blanket”. This gas envelope surrounds the impeller, causing it to rotate largely within gas rather than liquid. Furthermore, extremely high gas‑evolution rates may trigger broth overflow from the bioreactor.
Antifoam Agents
Most cell cultures produce foaming and foam‑stabilising substances such as proteins, polysaccharides and fatty acids. High gas flow rates and high agitation speeds exacerbate foam formation. Foaming gives rise to multiple operational challenges for reactors; hence foam control constitutes a critical consideration in bioreactor design. Foam escaping from the bioreactor provides entry routes for contaminating microorganisms and risks clogging off‑gas piping. When configuring bioreactors, 20 %‑30 % of total vessel volume must be reserved between the liquid surface and top headspace to accommodate foam layers and volume expansion caused by aeration.
Dosing specialised antifoam agents into the medium is a common practice for foam mitigation. Nevertheless, antifoams alter bubble surface chemistry and coalescence tendency and exert significant effects on kₗa. Most antifoams are potent surface‑tension reducers. Reduced surface tension decreases average bubble diameter and increases specific interfacial area a, which appears beneficial. However, it also lowers gas‑liquid interfacial mobility and reduces the liquid‑side mass‑transfer coefficient kₗ. For most silicone‑based antifoams, the reduction in kₗ usually outweighs the gain in a, leading to an overall drop in kₗa. Typical datasets for surface tension and kₗa versus antifoam concentration are presented in Figure 3A. Addition of merely 0.1 % antifoam reduced kₗa by nearly 50 %.
Oxygen Partial Pressure
According to standard international critical tables, the mole fraction of oxygen in air is 0.2099; for pure oxygen this value equals 1. At constant total pressure and temperature, sparging with pure oxygen instead of air increases oxygen solubility by 4.8‑fold. Alternatively, oxygen solubility can be elevated by sparging compressed air at higher total pressure. Both approaches, however, increase bioreactor operating costs.
A well‑performing fermenter is not equipment that simply “blows in more air”; rather, it efficiently transfers oxygen from gas phase to locations where cells can utilise it. This explains why straightforward replication of laboratory‑scale parameters frequently fails during fermentation scale‑up. Laboratory‑scale work focuses largely on strain performance, whereas industrial‑scale operations test the compatibility among microbial strains, equipment, processes and transport phenomena. For any aerobic fermentation system, the microbial strain defines production potential, while oxygen mass‑transfer determines whether such potential can be fully realised.
