Aeration Fundamentals: VVM and Superficial Gas Velocity
Bioreactor aeration scale-up depends on two parameters that behave very differently as vessel size increases: VVM and superficial gas velocity. Understanding the relationship between these two quantities is the first step to avoiding oxygen limitation—or impeller flooding—at production scale.
VVM (volume per volume per minute) is the volumetric gas flow rate divided by the liquid working volume. It is the most common way to express aeration rate in bioprocess protocols:
VVM = Q / VL
where Q is the gas flow rate in L/min and VL is the working volume in litres. Typical VVM ranges are 0.5–2.0 min−1 for microbial fermentation and 0.01–0.3 min−1 for mammalian cell culture.
Superficial gas velocity (vs) is the gas flow rate divided by the vessel cross-sectional area:
vs = Q / A = Q / (π DT² / 4)
where DT is the tank diameter. While VVM scales with volume (proportional to DT³), the cross-sectional area scales with DT². This means keeping VVM constant during scale-up causes vs to increase linearly with vessel diameter—a critical problem that causes flooding, excessive foaming, and poor gas dispersion at large scale.
The Van’t Riet Correlation for kLa Prediction
The Van’t Riet correlation (1979) remains the most widely used empirical equation for predicting kLa in stirred tank bioreactors. It relates kLa to the gassed power input per unit volume and superficial gas velocity using two sets of constants depending on the media’s coalescing behaviour.
The general form is:
kLa = K × (Pg/V)α × vsβ
| Parameter | Coalescent Media (water-like) | Non-Coalescent Media (salts, proteins) |
|---|---|---|
| K | 0.026 | 0.002 |
| α (Pg/V exponent) | 0.4 | 0.7 |
| β (vs exponent) | 0.5 | 0.2 |
| kLa units | s−1 | s−1 |
| Pg/V units | W/m3 | W/m3 |
| vs units | m/s | m/s |
The practical implication is important: in fermentation media, increasing agitation speed (which raises Pg/V) has a much larger effect on kLa than increasing air flow rate. This guides the DO cascade strategy discussed in Section 5.
Both exponents are less than 1, which means doubling either power input or gas velocity will less than double kLa. Oxygen transfer enhancement becomes progressively less efficient as you push harder—a fundamental limitation that shapes every aeration scale-up decision.
OTR & kLa Estimator
Calculate kLa for your bioreactor using Van’t Riet or Büchs correlations. Compare estimated kLa to organism oxygen demand.
Sparger Types and Selection Guide
Sparger selection directly determines bubble size distribution, which controls the interfacial area available for oxygen transfer. Smaller bubbles mean higher specific surface area and longer residence time—both of which increase kLa—but the trade-offs in foaming, shear, and CO2 stripping become critical at scale.
| Sparger Type | Pore/Hole Size | Bubble Diameter | Typical kLa (h−1) | Best Application |
|---|---|---|---|---|
| Open pipe | N/A (open end) | 5–10 mm | 30–60 | Simple systems, low shear |
| Drilled-hole (macro) | 0.5–1.0 mm | 3–6 mm | 50–80 | Cell culture (low shear) |
| Ring sparger | 0.2–0.5 mm | 1–3 mm | 80–150 | Large-scale (>10,000 L), CO2 stripping |
| Sintered metal | 20–40 µm | 0.5–1.0 mm | 150–250 | Microbial fermentation |
| Microsparger (porous frit) | 15–20 µm | 0.1–0.5 mm | 200–400 | High OUR, single-use bioreactors |
At scales above 2,000 L, CO2 removal becomes as important as O2 supply. Microspargers are highly efficient at dissolving oxygen but poor at stripping dissolved CO2 because their small bubbles saturate quickly. Ring spargers with 0.2–0.5 mm pores offer a compromise: adequate kLa for most mammalian cell culture processes with effective CO2 removal through larger, faster-rising bubbles.
A common approach at 2,000–10,000 L scale is dual sparging: a microsparger for O2/air delivery and a macro sparger (ring or drilled-hole) for CO2 stripping. This decouples the two mass transfer objectives.
Aeration Scale-Up Strategies Compared
No single aeration scale-up criterion works universally. The three most common strategies—constant VVM, constant vs, and constant kLa—each preserve different aspects of the gas-liquid environment, and the best choice depends on whether oxygen supply or CO2 removal is your primary constraint.
| Strategy | What Stays Constant | What Changes | Advantage | Limitation | Best For |
|---|---|---|---|---|---|
| Constant VVM | Gas volume per liquid volume per min | vs increases with DT | Simple; preserves CO2 stripping | Risk of flooding at large scale | Mammalian cell culture |
| Constant vs | Superficial gas velocity | VVM decreases with scale | Maintains gas holdup & hydrodynamics | CO2 removal may be inadequate | Microbial fermentation |
| Constant kLa | Volumetric mass transfer coefficient | Requires adjusting both P/V and vs | Directly preserves O2 transfer capacity | Hard to measure online; correlation uncertainty | O2-limited processes |
| Hybrid (P/V + VVM) | P/V at 1–3 W/L and minimum VVM | Fine-tuned per scale | Balances O2 and CO2; validated to 2,000 L | More complex to implement | mAb production, general purpose |
The hybrid approach deserves particular attention for mammalian cell culture at scale. By maintaining a minimum VVM (typically 0.01–0.05 min−1 for cell culture) to ensure CO2 stripping, while using constant P/V (1–3 W/L) to set the agitation-driven kLa, you address both gas transfer challenges simultaneously. This approach has been demonstrated from bench to pilot scale for multiple monoclonal antibody processes with consistent product quality attributes.
For more on how these criteria interact with impeller selection and tip speed, see our guide on 5 bioreactor scale-up criteria compared.
DO Cascade Control in Practice
A DO cascade controller activates aeration parameters in a defined sequence to maintain dissolved oxygen at the setpoint (typically 30–50% air saturation) while minimizing unnecessary shear, gas costs, and foaming. The standard cascade order is agitation first, then air flow, then oxygen enrichment, then backpressure.
The rationale for this sequence is based on efficiency and cell health:
- Agitation speed — First in the cascade because it has the largest effect on kLa in non-coalescent media (exponent 0.7 in Van’t Riet). Range: typically 100–300 rpm (lab) or 50–150 rpm (production).
- Air flow rate — Second because increasing vs has a smaller effect (exponent 0.2). Also aids CO2 stripping. Range: 0.5–2.0 VVM for microbial, 0.01–0.1 VVM for cell culture.
- Oxygen enrichment — Third because it raises the driving force (C* − CL) without changing hydrodynamics. Mix O2 into sparge gas from 21% up to 100%.
- Headspace pressure — Last resort because elevated pressure affects all dissolved gases (including CO2) and has safety and regulatory implications. Typically 0.2–0.5 bar overpressure maximum.
Gas Mixing Calculator
Calculate O2/air/N2/CO2 gas blends for target DO and pH control. Oxygen enrichment breakpoint analysis.
Common Aeration Scale-Up Problems
Most aeration scale-up failures trace back to one of four problems: impeller flooding, excessive foaming, cell damage from bubble bursting, or CO2 accumulation. All four become worse as bioreactor diameter increases.
Impeller Flooding
Impeller flooding is the condition where gas flow overwhelms the impeller’s ability to disperse bubbles. Air channels through the impeller zone without being broken into fine bubbles, resulting in poor kLa and a sudden drop in power draw (typically >30% reduction). For Rushton turbines, flooding occurs when the gas flow number exceeds a critical value:
Flcrit = Q / (N × Di³) ≈ 0.035
Monitor power draw in real time: a sharp drop indicates the onset of flooding. The fix is to increase impeller speed before increasing gas flow, and to ensure the lowest impeller sits within one impeller diameter above the sparger.
Foaming
Foaming intensifies with scale because higher total gas flow rates generate more bubbles at the surface. Protein-containing media and serum supplements are particularly prone to foaming. Silicone-based antifoam agents reduce kLa by 5–20%, so mechanical foam breakers or Pluronic F-68 (which protects cells without heavily suppressing kLa) are preferred in cell culture.
Cell Damage from Bubble Bursting
Bursting bubbles at the liquid surface—not the bubbles themselves in the bulk liquid—cause the majority of shear damage to mammalian cells. Smaller bubbles (from microspargers) concentrate more energy per burst event. Protective strategies include adding Pluronic F-68 at 0.1–0.2% (w/v), limiting tip speed below 1.5–2.0 m/s, and minimising unnecessary aeration.
CO2 Accumulation
At scales above 2,000 L, the headspace surface-area-to-volume ratio decreases, reducing passive CO2 removal. Dissolved CO2 above 120–150 mmHg inhibits growth and alters glycosylation in CHO cells. Maintain adequate VVM for stripping and consider a dedicated macro sparger for CO2 removal when using a microsparger for O2 supply. See also our article on dissolved oxygen control in CHO culture.
Worked Example: Scaling Aeration from 10 L to 5,000 L
This worked example demonstrates how to translate lab-scale aeration parameters to production scale using the Van’t Riet correlation and avoid common pitfalls like impeller flooding.
Worked Example — E. coli Fermentation Aeration Scale-Up
Lab scale (10 L): DT = 0.20 m, working volume = 7 L, 1.0 VVM, Pg/V = 3,000 W/m³, Rushton turbine
Step 1: Calculate lab-scale superficial gas velocity
Q = 1.0 VVM × 7 L = 7 L/min = 1.17 × 10−4 m³/s
A = π/4 × (0.20)² = 0.0314 m²
vs = 1.17 × 10−4 / 0.0314 = 3.7 × 10−3 m/s (3.7 mm/s)
Step 2: Calculate lab-scale kLa (non-coalescent media)
kLa = 0.002 × (3000)0.7 × (0.0037)0.2
kLa = 0.002 × 299.9 × 0.330
kLa = 0.198 s−1 = 713 h−1
Production scale (5,000 L): DT = 1.50 m, working volume = 3,500 L, target kLa ≥ 500 h−1
Step 3: Check what happens with constant VVM
Q = 1.0 VVM × 3500 L = 3500 L/min = 0.0583 m³/s
A = π/4 × (1.50)² = 1.767 m²
vs = 0.0583 / 1.767 = 0.033 m/s (33 mm/s)
⚠ vs = 33 mm/s — approaching the flooding limit! At 1.0 VVM the gas velocity is 9× higher than lab scale.
Step 4: Use constant vs instead and calculate required P/V
Target: vs = 0.010 m/s (10 mm/s, safe for production scale)
Q = 0.010 × 1.767 = 0.01767 m³/s = 1060 L/min
VVM = 1060 / 3500 = 0.30 VVM
To achieve kLa ≥ 500 h−1 (0.139 s−1):
0.139 = 0.002 × (Pg/V)0.7 × (0.010)0.2
0.139 = 0.002 × (Pg/V)0.7 × 0.398
(Pg/V)0.7 = 174.6
Pg/V = 174.6(1/0.7) = 2,370 W/m³ (2.4 W/L)
Result: At production scale, reduce VVM from 1.0 to 0.30 to maintain safe gas velocity, and use Pg/V of 2.4 W/L (vs. 3.0 W/L at lab) to achieve the target kLa. Total power draw = 2,370 × 3.5 m³ = 8.3 kW.
| Organism | Typical VVM (min−1) | Notes |
|---|---|---|
| E. coli | 1.0–2.0 | High OUR; 1.5 VVM typical for HCDC |
| S. cerevisiae | 1.0–3.0 | Up to 3 VVM for high-demand processes |
| Pichia pastoris | 1.0–2.0 | High OUR during methanol induction |
| CHO (ring sparger) | 0.01–0.30 | Total gas including overlay |
| CHO (microsparger) | 0.005–0.03 | Higher kLa per VVM; foaming risk |
| Insect cells (Sf9) | 0.01–0.05 | Shear-sensitive; low aeration rate |
| Filamentous fungi | 0.5–1.5 | High viscosity reduces transfer efficiency |
Scale-Up Calculator
Compare 5 scale-up criteria side-by-side: constant P/V, tip speed, Re, kLa, and mixing time. See exactly how aeration parameters change.
Frequently Asked Questions
What is VVM in bioreactors and how do you calculate it?
VVM (volume of gas per volume of liquid per minute) is the standard measure of aeration rate in bioreactors. Calculate it as VVM = Q / VL, where Q is the volumetric gas flow rate (L/min) and VL is the working liquid volume (L). Typical VVM ranges are 0.5–2.0 for microbial fermentation and 0.01–0.3 for mammalian cell culture.
How does kLa change when you scale up a bioreactor?
kLa typically decreases during scale-up because specific power input (P/V) is reduced to avoid excessive shear, and mixing becomes less uniform. Lab-scale kLa correlations systematically overestimate large-scale performance because intense turbulence is confined to the impeller zone. A 10 L reactor at 5 W/L may deliver kLa of 300 h−1, while a 10,000 L vessel at 1–2 W/L may only achieve 100–150 h−1.
What is the difference between VVM and superficial gas velocity?
VVM is gas flow rate divided by liquid volume (Q/VL), while superficial gas velocity (vs) is gas flow rate divided by the vessel cross-sectional area (Q/A). VVM scales with volume but vs scales with area. Keeping VVM constant during scale-up causes vs to increase proportionally with vessel diameter, which can cause impeller flooding and excessive foaming at production scale.
How do you prevent impeller flooding during bioreactor aeration?
Impeller flooding occurs when gas flow overwhelms the impeller’s ability to disperse bubbles. Prevent it by ensuring the impeller speed exceeds the minimum dispersion speed (NCD), keeping the gas flow number (Fl = Q / N×D3) below 0.035 for Rushton turbines, and using a lower impeller near the sparger for initial gas dispersion. Monitoring power draw drop (>30% indicates flooding) provides real-time detection.
What is the Van’t Riet correlation for kLa?
The Van’t Riet correlation (1979) predicts kLa as: kLa = K × (Pg/V)α × vsβ. For coalescent media (water-like): kLa = 0.026 × (Pg/V)0.4 × vs0.5. For non-coalescent media (salts/proteins): kLa = 0.002 × (Pg/V)0.7 × vs0.2. Units: kLa in s−1, Pg/V in W/m3, vs in m/s.
How do you control dissolved oxygen in a bioreactor?
Dissolved oxygen is controlled via a cascade strategy that activates parameters in sequence: (1) increase agitation speed to raise kLa, (2) increase air flow rate to raise vs, (3) enrich inlet gas with pure oxygen, (4) increase headspace pressure to raise DO saturation (C*). This cascade maintains DO at the setpoint (typically 30–50% air saturation) while minimising unnecessary shear and gas costs.
Related Tools
- OTR & kLa Estimator — Calculate kLa for stirred tanks and shake flasks using Van’t Riet and Büchs correlations
- Scale-Up Calculator — Compare 5 scale-up criteria side-by-side for any vessel geometry
- Gas Mixing Calculator — Calculate O2/air/N2/CO2 blends and oxygen enrichment breakpoints
References
- Van’t Riet K. Review of measuring methods and results in nonviscous gas-liquid mass transfer in stirred vessels. Ind Eng Chem Process Des Dev. 1979;18(3):357–364. doi:10.1021/i260071a001
- Xu S, et al. A practical approach in bioreactor scale-up and process transfer using a combination of constant P/V and vvm as the criterion. Biotechnol Prog. 2017;33(4):1146–1159. doi:10.1002/btpr.2489
- Garcia-Ochoa F, Gomez E. Bioreactor scale-up and oxygen transfer rate in microbial processes. Biotechnol Adv. 2009;27(2):153–176. doi:10.1016/j.biotechadv.2008.10.006
- Nienow AW. Reactor engineering in large scale animal cell culture. Cytotechnology. 2006;50(1):9–33. doi:10.1007/s10616-006-9005-8