How to Measure and Optimize Mixing Time for Bioreactor Scale-Up

May 2026 14 min read Bioprocess Engineering

Key Takeaways

Contents

  1. What Is Mixing Time and Why It Matters
  2. How to Measure Mixing Time: Three Proven Methods
  3. Mixing Time Correlations for Stirred-Tank Bioreactors
  4. Dimensionless Mixing Time and the Reynolds Number
  5. How Mixing Time Changes with Scale
  6. Strategies to Reduce Mixing Time at Large Scale
  7. Worked Example: Scaling from 2 L to 2,000 L
  8. Frequently Asked Questions

Mixing time is the single most revealing parameter for diagnosing concentration gradients inside a bioreactor. When a bioreactor scales from a 2 L bench vessel to a 20,000 L production tank, the mixing time can increase from 5 seconds to over 100 seconds, creating zones of depleted glucose, excess lactate, and pH excursions that directly reduce cell viability and product quality. This guide explains how to measure mixing time using industry-standard tracer methods, how to predict it from engineering correlations, and how to apply that knowledge during bioreactor scale-up.

Whether you are characterizing a new bioreactor, qualifying a scale-down model, or troubleshooting a production-scale process, understanding mixing time gives you a quantitative handle on something that otherwise remains invisible: how quickly your cells experience changes in their local environment.

What Is Mixing Time and Why It Matters

Mixing time (tm) is the time required to achieve a specified degree of homogeneity after introducing a perturbation into a stirred vessel. The industry standard is the 95% mixing time (t95), defined as the time from tracer injection until all measurement points remain within ±5% of the final equilibrium concentration.

In bioprocess engineering, mixing time directly governs:

Table 1. Typical mixing time ranges by bioreactor scale and organism type
Mixing time ranges at different bioreactor scales
Working Volume Typical tm (Mammalian) Typical tm (Microbial) Typical P/V
1-5 L3-8 s2-5 s10-50 W/m³ (mam.) / 0.5-3 kW/m³ (mic.)
10-50 L5-15 s3-8 s10-50 W/m³ / 0.5-3 kW/m³
100-500 L10-30 s5-12 s10-50 W/m³ / 0.5-2 kW/m³
1,000-5,000 L20-60 s8-20 s10-50 W/m³ / 0.5-2 kW/m³
5,000-25,000 L40-120 s15-40 s10-50 W/m³ / 0.3-1.5 kW/m³

These ranges assume standard baffled stirred-tank bioreactors with Rushton turbine or pitched-blade impellers at typical P/V values. Single-use bioreactors with bottom-mounted impellers often show 20-50% longer mixing times at equivalent P/V due to less efficient flow patterns.

How to Measure Mixing Time: Three Proven Methods

Three tracer-based methods dominate industrial practice, each with distinct advantages depending on the bioreactor type and measurement objective. All share the same basic protocol: inject a tracer at a defined location, monitor the response at one or more probe positions, and record the time to reach 95% of the final steady-state value.

Mixing Time Measurement Methods pH Tracer NaOH pH 1 pH 2 Inject acid or base Monitor pH response at multiple positions ✓ Probes already in vessel ✓ Non-invasive ✓ Works in SUS & SS ✗ Probe response lag ✗ Can affect cell culture Precision: ±10-15% Conductivity Tracer NaCl C1 C2 Inject salt solution Track conductivity response over time ✓ Fastest response time ✓ Best precision ✓ Quantitative output ✗ Requires extra probes ✗ Salt alters osmolality Precision: ±5-8% Colorimetric Camera Acid Base pH indicator dyes Record colour change via camera + image analysis ✓ Whole-field visualization ✓ No probes needed ✓ Ideal for SUS/transparent ✗ Subjective endpoint ✗ Opaque vessels excluded Precision: ±15-25%
Figure 1. Three standard mixing time measurement methods. The pH tracer method is most common in industry because pH probes are already installed. Conductivity offers the best precision. Colorimetric is ideal for transparent single-use systems.
Diagram comparing three mixing time measurement methods: pH tracer using acid or base injection monitored by pH probes, conductivity tracer using salt solution monitored by conductivity probes, and colorimetric method using pH indicator dyes recorded by camera. Each method shows its advantages, disadvantages, and measurement precision.

pH Pulse Tracer Method

The pH pulse method is the most widely used approach in industrial bioreactors because it requires no additional instrumentation. Inject a small volume of concentrated acid (typically 0.5-1.0 M HCl) or base (0.5-1.0 M NaOH) near the liquid surface and monitor the pH response at two or more probe locations. The tracer volume should be small enough to cause a measurable perturbation (0.1-0.3 pH units) without significantly altering the bulk pH.

Position probes at locations representing the worst-case mixing zones: one near the impeller discharge and one at the furthest point from the impeller, typically near the liquid surface in the baffle region. The mixing time is the longest time recorded across all probe positions.

Conductivity Tracer Method

Injecting a pulse of concentrated NaCl solution (1-5 M) and tracking the conductivity response provides the fastest sensor response time and highest precision (±5-8%). Conductivity probes have response times of 0.1-0.5 seconds, compared to 3-10 seconds for glass pH electrodes, making this method essential when mixing times are short (under 10 s).

Colorimetric Decolorization Method

The dual-indicator system for mixing time (DISMT) uses two pH indicators (typically bromothymol blue and methyl red) added to the vessel. A colour change is triggered by acid or base addition, and the time until uniform colour is recorded by video camera and quantified through image analysis. This method is particularly valuable for single-use bioreactors where inserting additional probes is impractical.

Mixing Time Correlations for Stirred-Tank Bioreactors

In the fully turbulent regime (Re > 10,000), two correlations dominate the bioprocessing literature. Both assume standard geometry: baffled vessel, H/T ≈ 1, single impeller, D/T = 0.25-0.5.

The Grenville-Nienow Correlation

The most widely cited mixing time correlation for single-impeller stirred tanks in turbulent flow is:

Grenville-Nienow Correlation (Turbulent, Re > 10,000)

N · tm = 5.9 · (T / D)2 · Po−1/3

Where:

The product N · tm is the dimensionless mixing time (also called the mixing number, Nt). It represents the number of impeller revolutions needed to achieve 95% homogeneity. A key insight from this correlation is that all impeller types of equal D/T ratio are equally energy-efficient at blending in the turbulent regime, because the Po−1/3 dependency is weak.

The Power-per-Volume Form

Rearranging the Grenville-Nienow correlation in terms of specific power input P/V gives:

P/V Form of the Mixing Time Correlation

tm = 5.9 · (T / D)2 · Po−1/3 · (ρ · Po / (P/V))1/3 · (4 / π)1/3 · T−2/3

Which simplifies to the widely used engineering approximation:

tm ∝ (P/V)−1/3 · T2/3

This proportionality reveals the fundamental scale-up challenge: at constant P/V, mixing time increases as the two-thirds power of tank diameter. Doubling T increases mixing time by a factor of 22/3 ≈ 1.59.

Table 2. Dimensionless mixing time values for common impeller types (turbulent regime, D/T = 0.33, single impeller)
Dimensionless mixing time N · tm by impeller type
Impeller Type Power Number (Po) N · tm Flow Pattern
Rushton turbine (6-blade)5.032Radial
Smith turbine (6-blade)3.237Radial
Pitched-blade turbine (4-blade, 45°)1.350Axial-radial
Hydrofoil (Lightnin A315)0.381Axial
Marine propeller0.3577Axial

Lower-power impellers require more revolutions to mix but consume less energy per revolution. The net energy to achieve mixing (E = P · tm) is approximately constant across impeller types, which is why the choice between Rushton and hydrofoil impellers for mixing is driven by shear sensitivity and gas dispersion requirements rather than mixing efficiency alone.

Dimensionless Mixing Time and the Reynolds Number

The dimensionless mixing time N · tm is constant in the fully turbulent regime (Re > 10,000) but rises sharply in the transitional (10 < Re < 10,000) and laminar (Re < 10) regimes. This behavior is critical for bioreactors processing viscous broths, particularly fungal fermentations and high-cell-density mammalian cultures.

The impeller Reynolds number is defined as:

Re = ρ · N · D2 / μ

Where ρ is the fluid density (kg/m³), N is the impeller speed (rev/s), D is the impeller diameter (m), and μ is the dynamic viscosity (Pa·s).

Figure 2. Dimensionless mixing time (N · tm) versus impeller Reynolds number for four impeller types. In the turbulent regime (Re > 10,000), N · tm reaches a constant value. Below Re 1,000, mixing time increases sharply as viscous forces dominate.

For most aqueous bioprocess media (μ ≈ 0.001 Pa·s), bioreactors operate well into the turbulent regime even at moderate impeller speeds. A 2 L vessel with a 5 cm Rushton turbine at 300 RPM gives Re ≈ 12,500. However, viscous fungal broths (μ = 0.01-0.1 Pa·s) can drop the Reynolds number into the transitional zone, where mixing times increase by 2-5 fold compared to turbulent predictions.

How Mixing Time Changes with Scale

Mixing time increases with bioreactor volume at every practical scale-up criterion. The magnitude of the increase depends on which parameter is held constant during scale-up.

Table 3. Mixing time increase factors for a 1,000-fold volume scale-up (2 L to 2,000 L) under different scale-up criteria
Effect of scale-up criterion on mixing time
Scale-Up Criterion Held Constant Mixing Time Factor RPM at 2,000 L Notes
Constant P/V×4.6~170Most common; preserves kLa approximately
Constant tip speed×6.8~130Preserves max shear; longer mixing time
Constant Re×10~80Very slow mixing; rarely used alone
Constant mixing time×1.0~1,800Impractical: extreme P/V, severe cell damage

Constant P/V is the most common criterion because it approximately preserves both kLa (oxygen transfer) and bulk fluid velocities. The trade-off is a 4-6 fold increase in mixing time. For mammalian cell culture where mixing times at bench scale are 5-8 seconds, this translates to 25-45 seconds at 2,000 L, which is generally acceptable. For microbial fermentations where the target mixing time is under 10 seconds, a 5-fold increase can create significant substrate gradients.

Figure 3. Predicted mixing time versus bioreactor working volume at constant P/V = 50 W/m³ (mammalian) and constant P/V = 1.5 kW/m³ (microbial). Both curves follow the tm ∝ V2/9 relationship derived from the Grenville-Nienow correlation. Dashed lines indicate typical upper limits for acceptable mixing.

Strategies to Reduce Mixing Time at Large Scale

When predicted mixing times exceed acceptable thresholds, several engineering strategies can improve mixing without increasing shear stress to damaging levels.

  1. Multiple impellers — Adding a second or third impeller on the same shaft reduces mixing time by 30-50% compared to a single impeller at the same total power input. The optimal spacing between impellers is 1.0-1.5 times the impeller diameter. For H/T > 1.5, multiple impellers are essentially mandatory.
  2. Increase D/T ratio — Increasing the impeller-to-tank diameter ratio from D/T = 0.33 to D/T = 0.5 reduces the dimensionless mixing time N · tm by a factor of (0.33/0.5)2 ≈ 2.3. This is one of the most effective interventions at any scale.
  3. Axial-radial impeller combinations — Pairing a lower radial-flow impeller (Rushton for gas dispersion) with an upper axial-flow impeller (pitched blade or hydrofoil for top-to-bottom circulation) produces shorter mixing times than dual Rushton configurations at equivalent power input.
  4. Bottom-entry feed addition — Adding concentrated feeds near the impeller discharge zone rather than at the liquid surface reduces the effective mixing distance for the most critical component.
  5. Increase power input selectively — A modest P/V increase of 50% reduces mixing time by only (1.5)−1/3 ≈ 13%, so this approach alone is rarely sufficient. It is most effective combined with geometric changes.

Scale-Up Calculator

Compare P/V, tip speed, kLa, Re, and mixing time across scales. Input your bench-scale parameters and see how each criterion translates to production.

Open Calculator

Worked Example: Scaling Mixing Time from 2 L to 2,000 L

This worked example predicts the 95% mixing time at production scale (2,000 L) given bench-scale measurements and a constant P/V scale-up strategy.

Worked Example — Mixing Time Scale-Up

Given (bench scale, 2 L):

Step 1: Verify with correlation

N · tm = 5.0 × 6 = 30
Predicted: 5.9 × (0.13/0.043)2 × 5.0−1/3
= 5.9 × 9.14 × 0.585 = 31.5

Measured (30) agrees with predicted (31.6) within 5%. Good confidence in the correlation for this system.

Step 2: Calculate P/V at bench scale

P = Po × ρ × N3 × D5
= 5.0 × 1000 × 5.03 × 0.0435
= 5.0 × 1000 × 125 × 1.47 × 10−7
= 0.092 W
P/V = 0.092 / 0.002 = 46 W/m³

Step 3: Determine production-scale geometry and RPM

V2 = 2,000 L = 2.0 m³ (scale factor = 1,000×)
T2 = T1 × (V2/V1)1/3 = 0.13 × 10 = 1.30 m
D2 = D/T × T2 = 0.33 × 1.30 = 0.429 m

For constant P/V (46 W/m³):
N2 = (P/V × V2 / (Po × ρ × D25))1/3
= (46 × 2.0 / (5.0 × 1000 × 0.4295))1/3
= (92 / (5.0 × 1000 × 0.01454))1/3
= (92 / 72.7)1/3
= 1.2651/3 = 1.082 rev/s = 65 RPM

Step 4: Predict mixing time at production scale

N · tm = 5.9 × (1.30/0.429)2 × 5.0−1/3
= 5.9 × 9.18 × 0.585 = 31.7
tm,2 = 31.7 / 1.082 = 29.3 s ≈ 29 s

Result: Mixing time increases from 6 s to ~29 s (a factor of 4.9) when scaling from 2 L to 2,000 L at constant P/V. This is consistent with the (V2/V1)2/9 = 10000.222 = 4.6 approximation. For a mammalian cell culture process, 29 s is within the acceptable range. For a microbial fermentation, additional measures (multiple impellers or increased D/T) would be warranted.

Heat Transfer Calculator

Mixing quality directly affects heat transfer at large scale. Calculate jacket and coil heat transfer coefficients to ensure temperature uniformity.

Open Calculator

When to Worry About Mixing Time

Production-scale tm > 60 s? (Mammalian cell culture) Yes Action Required pH/DO gradients likely significant Add 2nd impeller Reduces tm 30-50% Increase D/T ratio From 0.33 to 0.45-0.50 No Microbial? tm > 15 s? (E. coli, yeast, Pichia) Yes Monitor gradients Scale-down model study No Acceptable Proceed with scale-up Always validate with scale-down model studies Two-compartment or multi-compartment systems simulate large-scale gradients at bench scale (Paul et al. 2020)
Figure 4. Decision tree for evaluating mixing time impact at production scale. Mammalian cell culture thresholds are more forgiving than microbial processes due to lower metabolic rates.
Decision tree starting with the question: is production-scale mixing time greater than 60 seconds for mammalian cell culture? If yes, action is required (add second impeller or increase D/T ratio). If no, check if the process is microbial with mixing time greater than 15 seconds, which warrants gradient monitoring via scale-down models.

OTR/kLa Estimator

Mixing time and kLa are coupled. Poor mixing reduces effective kLa. Estimate oxygen transfer rates and identify DO limitations.

Open Calculator

Frequently Asked Questions

What is mixing time in a bioreactor?

Mixing time (tm) is the time required to achieve a specified degree of homogeneity, typically 95%, after introducing a tracer into a bioreactor. It is measured from the moment of tracer addition until the concentration at one or more probe locations remains within 5% of the final equilibrium value. Typical mixing times range from 3-10 seconds at bench scale (1-10 L) to 30-120 seconds at production scale (5,000-25,000 L).

How do you measure mixing time in a bioreactor?

The three most common methods are pH tracer (inject acid or base pulse, monitor with pH probes), conductivity tracer (inject salt solution, track conductivity), and colorimetric decolorization (add pH indicator dyes, record the time until uniform colour). The pH tracer method is preferred in industry because pH probes are already installed in most bioreactors, making the measurement non-invasive and easy to implement.

What is the mixing time correlation for stirred-tank bioreactors?

The widely used Grenville-Nienow correlation for turbulent flow (Re > 10,000) is N · tm = 5.9 (T/D)2 Po−1/3, where N is impeller speed (rev/s), tm is 95% mixing time (s), T is tank diameter (m), D is impeller diameter (m), and Po is the power number. This predicts that all impeller types of equal D/T ratio are equally energy-efficient at achieving homogeneity in the turbulent regime.

Why does mixing time increase with bioreactor scale?

Mixing time scales approximately as tm ∝ (P/V)−1/3 · T2/3. At constant P/V, increasing tank diameter T causes mixing time to rise because the liquid must travel a greater distance from the impeller to the vessel walls and dead zones. A 1,000-fold volume increase (e.g. 2 L to 2,000 L) at constant P/V typically increases mixing time by a factor of 4-6.

What is an acceptable mixing time for mammalian cell culture?

For mammalian cell culture bioreactors, mixing times below 30 seconds at bench scale and below 60-120 seconds at production scale are generally considered acceptable. Beyond 120 seconds, pH and dissolved oxygen gradients become significant enough to affect cell growth and product quality. For microbial fermentations, mixing times should ideally stay below 10-15 seconds because faster metabolic rates amplify the impact of concentration gradients.

Related Tools

References

  1. Nienow, A.W. (1997). On impeller circulation and mixing effectiveness in the turbulent flow regime. Chemical Engineering Science, 52(15), 2557-2565. doi:10.1016/S0009-2509(97)00072-9
  2. Grenville, R.K. & Nienow, A.W. (2003). Blending of miscible liquids. In: Paul, E.L., Atiemo-Obeng, V.A. & Kresta, S.M. (eds.) Handbook of Industrial Mixing: Science and Practice, Wiley, pp. 507-542. doi:10.1002/0471451452.ch9
  3. Cabaret, F., Bonnot, S., Fradette, L. & Tanguy, P.A. (2007). Mixing time analysis using colorimetric methods and image processing. Industrial & Engineering Chemistry Research, 46(14), 5032-5042. doi:10.1021/ie0613265
  4. Zakrzewski, R., Lee, K. & Lye, G.J. (2022). Development of a miniature bioreactor model to study the impact of pH and DOT fluctuations on CHO cell culture performance. Biotechnology Progress, 38(6), e3264. doi:10.1002/btpr.3264
  5. Kaiser, S.C., Werner, S., Jossen, V., Kraume, M. & Eibl, D. (2017). Development of a method for reliable power input measurements in conventional and single-use stirred bioreactors at laboratory scale. Engineering in Life Sciences, 17(5), 500-511. doi:10.1002/elsc.201600096

Resources & Further Reading