The many types of bioreactors in use today look bewildering as a list, but they collapse into a small number of families once you ask a single question: how does mechanical energy get into the vessel? That one distinction sets the power input, the shear field, the oxygen transfer capacity and the practical scale ceiling of every design. This guide walks through ten bioreactor types, gives the quantitative range that separates them, and ends with a decision path and a comparison table you can work through for your own process.
How bioreactors are classified
Bioreactors are classified by their energy input mechanism, and everything else follows from it. There are three groups. Mechanically agitated reactors use a rotating impeller. Pneumatically agitated reactors use the buoyancy of injected gas. Externally moved or unagitated reactors either rock the whole vessel or rely on flow through a stationary bed.
This matters because the group determines which operating variables you can set independently. In a stirred tank, agitation and gas flow are separate knobs. In an airlift, one gas flow sets mixing, oxygen transfer and shear simultaneously. In a packed bed, there is no agitation at all and mass transfer depends entirely on flow through the bed and diffusion into the particle.
Bioreactors divide into three groups. Mechanically agitated reactors such as stirred tanks deliver energy through an impeller at 0.1 to 10 kilowatts per cubic metre and allow mixing and aeration to be set independently. Pneumatically agitated reactors such as airlift and bubble column deliver energy through injected gas, where power input equals liquid density times gravity times superficial gas velocity, and a single gas flow controls mixing, oxygen transfer and shear together. Moved or unagitated reactors such as wave, packed-bed, fluidised-bed, tray, membrane and photobioreactors have the lowest shear and are limited by flow and diffusion into particles.
Stirred-tank bioreactors
The stirred-tank bioreactor is the industry default because it is the only design that decouples mixing from aeration. An impeller supplies the mixing energy and a sparger supplies the gas, so the two can be adjusted independently. That single property is why the stirred tank survives in almost every sector despite being neither the gentlest nor the cheapest option.
Power input follows the standard relation P = Np ρ N³ D⁵, where Np is the impeller power number, N the rotational speed in s-1, D the impeller diameter in metres and ρ the broth density. Because power scales with the fifth power of impeller diameter and the cube of speed, small geometry changes have large consequences. Bacterial and yeast fermentations typically run at 1–10 kW/m³; mammalian cell culture runs one to two orders of magnitude lower, and microcarrier culture lower still.
Impeller choice sets the character of the reactor. A Rushton turbine (Np ≈ 5) is a radial-flow, high-shear, gas-dispersing design suited to oxygen-demanding microbial work. A pitched-blade turbine (Np ≈ 1.3) pumps axially and is gentler. Marine and elephant-ear impellers (Np ≈ 0.3–0.5) are gentler still and dominate in cell culture.
Scale-Up Calculator
Match P/V, tip speed or kLa between two vessel sizes and see which criterion breaks first.
Airlift and bubble column bioreactors
Pneumatically agitated bioreactors replace the impeller with injected gas, so there is no rotating shaft, no shaft seal and no high-shear impeller zone. A bubble column is the simplest form: gas is sparged into the base of a vertical vessel and rises, mixing the liquid as it goes. An airlift bioreactor adds a draft tube or an external loop that separates the vessel into a gassed riser and a less-gassed downcomer. The density difference drives an orderly circulation rather than the random churn of a bubble column.
Specific power input in a pneumatic reactor is simply P/V = ρ g uG, where uG is the superficial gas velocity. There is only one knob. Raising oxygen transfer means raising gas flow, which also raises mixing intensity, shear at the sparger and foaming tendency. This is the central trade against a stirred tank.
Airlift reactors are strong where sterility over long campaigns matters, since the shaft seal is a classic contamination route, and where the culture is shear-sensitive. They lose out as viscosity rises: circulation slows, gas holdup falls, and the reactor can effectively stall.
Worked example — stirred tank vs airlift at the same scale
A 1 m³ working volume, water-like broth (ρ = 1000 kg/m³).
Stirred tank, Rushton Np = 5.0, D = 0.30 m, N = 200 rpm = 3.333 s-1
P = 5.0 × 1000 × (3.333)³ × (0.30)⁵
= 5.0 × 1000 × 37.04 × 0.00243 = 450 W
P/V = 450 W / 1 m³ = 450 W/m³
Airlift, superficial gas velocity uG = 0.02 m/s
P/V = ρ g uG = 1000 × 9.81 × 0.02 = 196 W/m³
The stirred tank delivers about 2.3× the average power input. The more important difference is not the average but the distribution: in the stirred tank almost all of that energy dissipates in the small volume swept by the impeller, so the maximum local energy dissipation is far higher than the average. In the airlift it is spread through the riser. That is why airlift reactors are chosen for shear-sensitive cultures even when average P/V is comparable.
Packed-bed, fluidised-bed and immobilised-cell designs
Immobilising the biocatalyst changes the reactor question entirely, because the rate-limiting step moves from bulk mixing to diffusion into the particle. Immobilised-cell systems are run in five configurations: stirred-tank, fixed-bed (packed-bed and trickle-bed), fluidised-bed, gas-agitated, and membrane.
In a packed-bed bioreactor the particles are stationary and liquid flows through, approximating plug flow. Plug flow avoids back-mixing, so conversion per pass is higher than in a well-mixed vessel. The costs are channelling, pressure drop, and clogging as biomass accumulates. A trickle-bed is the same geometry with liquid flowing down over a gas-filled bed.
In a fluidised-bed bioreactor the upward liquid velocity suspends the particles, expanding the bed. Mass transfer improves and clogging largely disappears, at the price of particle attrition and a more demanding retention system.
Performance in every immobilised system is captured by the effectiveness factor η, defined as the ratio of the substrate consumption rate with diffusional resistance to the rate without it. An η of 1 means the particle is fully accessible; a low η means the core is starved and you are paying for biocatalyst that is doing nothing. Reducing particle diameter raises η but also raises pressure drop, which is the fundamental design tension in fixed beds.
Membrane bioreactors
A membrane bioreactor integrates reaction and separation in one unit. The membrane retains the biocatalyst while permitting product or spent medium to leave, so cells reach very high densities and the product is continuously removed. That matters most when the product inhibits its own formation.
Hollow-fibre formats give an extremely high surface-area-to-volume ratio and are widely used for high-density mammalian culture and for perfusion-style operation. The recurring failure mode is fouling: as the biomass layer thickens, transmembrane flux falls and nutrient gradients develop along the fibre.
Wave, rocking and single-use formats
Wave and rocking bioreactors move the vessel rather than anything inside it. A pre-sterilised plastic bag sits on a rocking platform; the wave generated at the liquid surface provides both mixing and oxygen transfer. There is no impeller, no sparger and no shaft, so shear is the lowest of any actively mixed design and the contamination risk is very low.
Single-use format is a separate axis from reactor type, not a type in itself. Stirred, orbitally shaken and wave-mixed single-use reactors all exist, and each scales differently. Rocking-motion systems in cell therapy run from roughly 0.3 L to 500 L, pneumatically driven single-use vessels from about 3 L to 500 L, packed and fluidised-bed single-use formats stay at 5 L or below, while stirred single-use systems span 0.015 L to 1000 L. That spread is the practical reason stirred tanks dominate: only they cover the whole range in one family.
| Format | Working volume (L) | Energy input |
|---|---|---|
| Rocking motion | 0.3–500 | Vessel movement |
| Pneumatically driven | 3–500 | Injected gas |
| Packed and fluidised bed | ≤ 5 | Liquid flow |
| Stirred tank | 0.015–1000 | Impeller |
Photobioreactors
A photobioreactor adds light as a substrate, and light is unlike every other nutrient because it cannot be mixed into the broth. It arrives at a surface and is attenuated with depth, so the design problem is geometric: maximise illuminated surface area per unit volume without making the vessel impractical.
That constraint produces the characteristic forms — flat panels, tubular loops, and open raceway ponds. As cell density rises, self-shading means the culture becomes light-limited at a depth of a few centimetres, which is why closed photobioreactors use narrow light paths and why productivity is usually quoted per unit area rather than per unit volume.
Solid-state fermentation bioreactors
Solid-state fermentation grows organisms on a moist solid substrate with little or no free water, and its reactors are limited by heat removal rather than by oxygen supply. Without bulk water to carry metabolic heat away, steep temperature and gas-concentration gradients form within the bed.
Tray bioreactors solve this by keeping every bed thin enough that heat escapes by conduction and natural convection. The consequence is that scale-up means adding trays, not deepening them, which is labour-intensive but predictable. Packed-bed designs force air through the bed to remove heat convectively. Rotating-drum and mixed designs agitate the bed to break gradients up.
Mixing is the crux. Agitating the bed helps overcome the spatial temperature gradients that come with convective heat removal, but whether you can use it depends entirely on how sensitive the organism is to being tumbled. Filamentous fungi bind particles together, and the contractive force of that binding, combined with bed compaction, causes channelling that short-circuits the airflow and defeats the cooling.
Microbioreactors
Microbioreactors compress a controlled cultivation into millilitre or even nanolitre volumes so that many conditions can run in parallel. They span micro-Petri dishes, instrumented microtiter plates, microfluidic chips with culture chambers, and miniature stirred vessels of roughly 10–250 mL with real pH and dissolved-oxygen control.
The extreme end is genuinely small: one early continuously operated microbioreactor ran six units of 16 nL working volume in parallel on a single chip, with active suppression of biofilm formation so that populations could be tracked for hundreds of hours. Instrumented millilitre-scale systems are now routine for parallel process development.
The recurring caution applies to all of them. A screening format only earns its place if its engineering environment resembles the one you will scale into. If a microtiter plate is oxygen-limited or mixing out of phase, it will rank your candidates by how well they tolerate that artefact.
OTR & kLa Estimator
Check whether a flask, plate or stirred tank can actually meet your culture's oxygen demand before you run it.
How to choose the right type of bioreactor
Four questions, in order, narrow the field faster than any feature comparison. Answer them before you look at a single vendor page.
- Is the biocatalyst free or immobilised? Immobilisation points immediately to fixed-bed, fluidised-bed or membrane, and brings the effectiveness factor into play.
- How shear-sensitive is it? This decides between mechanical and pneumatic agitation. Compare the Kolmogorov microscale against the size of the entity you are protecting, not against intuition.
- What is the oxygen demand? A high, time-varying demand needs independent control of agitation and aeration, which only the stirred tank offers.
- Is the substrate liquid or a wet solid? A solid substrate rules out every submerged design and puts you in the solid-state family, where heat removal becomes the design driver.
Decision tree. First ask whether the substrate is solid or liquid. A solid substrate leads to the solid-state family: tray, packed-bed or rotating drum. A liquid substrate leads to asking whether the biocatalyst is immobilised. If immobilised, choose packed-bed, fluidised-bed or membrane and watch the effectiveness factor. If not immobilised, ask whether the culture is shear-sensitive. If shear-sensitive, choose airlift or bubble column, or a wave reactor when oxygen demand is low. If not shear-sensitive, choose a stirred tank and match the impeller to the shear tolerance. Scale and single-use format follow from these answers rather than driving them.
Side-by-side comparison
The table below collects the operating characteristics that actually differentiate the families. Power inputs are order-of-magnitude working ranges rather than hard limits, since every one depends on geometry and broth properties.
| Type | Energy input | Typical P/V | Relative shear | Independent mixing & aeration? | Typical use |
|---|---|---|---|---|---|
| Stirred-tank | Impeller | 0.1–10 kW/m³ | High (impeller zone) | Yes | Microbial and mammalian production, the default |
| Airlift | Injected gas | 0.05–0.5 kW/m³ | Low | No | Shear-sensitive cultures, long sterile campaigns |
| Bubble column | Injected gas | 0.05–0.5 kW/m³ | Low | No | Simple aerobic processes, wastewater |
| Packed-bed | Liquid flow | Very low | Very low | No | Immobilised enzyme or cell conversion |
| Fluidised-bed | Liquid flow | Low | Low–moderate | No | Immobilised cells where clogging is a risk |
| Membrane | Pump / flow | Low | Low | Partly | High-density culture, in-situ product removal |
| Wave / rocking | Vessel motion | Very low | Lowest of actively mixed | No | Seed train, cell therapy, low O₂ demand |
| Photobioreactor | Gas and/or pump | Low | Low | No | Algae and cyanobacteria; light-limited |
| Solid-state (tray, bed, drum) | None to low | Very low | Low (bed damage instead) | No | Fungal enzymes, koji, biopesticides |
| Microbioreactor | Varies | Format-dependent | Format-dependent | Sometimes | Parallel screening and process development |
Frequently asked questions
What are the main types of bioreactors?
The main types are stirred-tank, pneumatically agitated (airlift and bubble column), fixed-bed (packed-bed and trickle-bed), fluidised-bed, membrane, wave or rocking, photobioreactor, solid-state fermentation reactors, and microbioreactors. Grouping them by how mechanical energy enters the vessel — impeller, injected gas, or movement of the vessel itself — explains most of their differences in power input, shear and scale.
Why is the stirred-tank bioreactor the industry default?
Because it decouples mixing from aeration. Agitation and gas flow are independent knobs, so oxygen transfer can be raised without raising gas rate, and mixing can be raised without changing sparging. It also spans the widest scale range in one family, from roughly 15 mL microbioreactor vessels to 1000 L and beyond, so a process can stay in the same reactor type from screening to manufacturing.
When should you use an airlift bioreactor instead of a stirred tank?
When the culture is shear-sensitive, when sterility over a long campaign matters, or when the broth is low-viscosity. There is no shaft seal, which is a common contamination route, and no impeller zone, so maximum local energy dissipation is much lower. The trade-off is that a single gas flow sets mixing, oxygen transfer and shear together, and performance degrades sharply as viscosity rises.
What limits the scale of a solid-state fermentation bioreactor?
Heat removal, not oxygen. With little free water to carry heat away, metabolic heat accumulates and steep temperature and gas gradients form in the bed. Tray reactors keep every bed thin, so scale-up means more trays rather than deeper beds. Packed-bed and drum designs move heat by forced aeration or mixing, but mixing damages fungal mycelium and can compact the bed into channels.
What is the difference between a packed-bed and a fluidised-bed bioreactor?
In a packed bed the particles are stationary and liquid flows through in near plug flow, giving high conversion per pass but risking channelling and clogging. In a fluidised bed the upward velocity suspends the particles, so the bed expands and behaves closer to a well-mixed vessel, avoiding clogging and improving mass transfer at the cost of particle attrition.
How do you choose the right type of bioreactor?
Ask four questions in order: is the biocatalyst free or immobilised, how shear-sensitive is it, what is the oxygen demand, and is the substrate liquid or a wet solid. Those four answers eliminate most families. Scale and single-use format are usually consequences rather than independent choices.
Related tools
- Scale-Up Calculator — compare P/V, tip speed and kLa criteria between two vessel sizes.
- OTR & kLa Estimator — check whether a given vessel can meet your oxygen demand.
- Microcarrier Planner — for adherent culture in a stirred tank, including the agitation window.
References
- Zhong J-J. (2011). Bioreactor Engineering. In: Comprehensive Biotechnology (2nd ed.), Academic Press, pp. 165–177.
- Garcia-Ochoa F, Santos VE, Gomez E. (2019). Stirred Tank Bioreactors. In: Comprehensive Biotechnology (3rd ed.), Elsevier.
- Guieysse B, Quijano G, Muñoz R. (2011). Airlift Bioreactors. In: Comprehensive Biotechnology (2nd ed.), Academic Press, pp. 199–212.
- Mitchell DA, de Lima Luz LF, Krieger N, Berovic M. (2011). Bioreactors for Solid-State Fermentation. In: Comprehensive Biotechnology (2nd ed.), Academic Press, pp. 347–360.
- Mandenius C-F (ed.). (2016). Bioreactors: Design, Operation and Novel Applications. Wiley-VCH, Weinheim. doi:10.1002/9783527683369