A chemostat (chemical environment is static) is a continuous culture device that maintains microbial cells at a constant, user-defined growth rate by feeding fresh medium at a fixed flow rate and removing culture at the same rate. Invented independently by Jacques Monod and by Novick and Szilard in 1950, the chemostat remains the gold standard for studying growth-rate-dependent physiology, adaptive evolution, and continuous bioprocessing. This guide derives the chemostat equations from first principles, walks through a worked example with real E. coli parameters, and compares the chemostat to turbidostat and auxostat alternatives.
What Is a Chemostat?
A chemostat is a stirred bioreactor with a constant-rate feed pump and an overflow outlet that keeps the culture volume fixed. Fresh sterile medium enters at volumetric flow rate F (mL/h or L/h), the vessel has a fixed working volume V (mL or L), and spent culture plus cells exits through the overflow at the same rate F. Because one nutrient in the feed is deliberately supplied below excess, that nutrient limits growth and the culture self-regulates to a steady state where the specific growth rate equals the dilution rate.
The key insight is control: by turning a pump dial, you set the growth rate of every cell in the vessel. No other culture system offers this. Batch cultures have a constantly changing growth rate (lag, exponential, stationary). Fed-batch cultures slow growth as nutrients accumulate or deplete. Only the chemostat holds μ constant for as long as you supply medium.
Dilution Rate: The Master Control Variable
Dilution rate is the single most important parameter in continuous culture. It is defined as the volumetric flow rate of fresh medium divided by the culture volume:
The reciprocal of D is the residence time τ = V/F = 1/D, representing the average time a molecule of medium or a cell spends in the vessel. At D = 0.2 h−1, the residence time is 5 hours, meaning the entire vessel volume is replaced every 5 hours.
In a well-mixed chemostat at steady state, cells that grow faster than D accumulate until they consume the limiting nutrient down to a level that slows growth to exactly D. Cells that grow slower than D are washed out. The net result is a self-regulating equilibrium: μ = D.
This relationship gives the chemostat its power. Want to study how E. coli behaves at μ = 0.1 h−1 versus μ = 0.5 h−1? Set D to those values and wait for steady state. No other culture system lets you decouple growth rate from time so cleanly.
| D (h−1) | τ = 1/D (h) | μ/μmax | Typical Application |
|---|---|---|---|
| 0.05 | 20.0 | 7% | Starvation physiology, stringent response studies |
| 0.10 | 10.0 | 14% | Maintenance metabolism, quorum sensing |
| 0.20 | 5.0 | 29% | Carbon-limited steady state, proteomics |
| 0.30 | 3.3 | 43% | Adaptive evolution at moderate growth |
| 0.50 | 2.0 | 71% | High-rate continuous production |
| 0.60 | 1.7 | 86% | Near-maximum productivity zone |
| 0.70 | 1.4 | 100% | Washout boundary (Dcrit) |
Steady-State Equations: Biomass and Substrate
The steady-state concentrations of biomass and substrate in a chemostat follow directly from the Monod growth model and mass balances. At steady state, dX/dt = 0 and dS/dt = 0, yielding two equations that every bioprocess engineer should know.
The mass balance on biomass in the vessel is:
At steady state dX/dt = 0. For a non-trivial solution (X ≠ 0), this requires μ = D.
Substituting the Monod equation μ = μmax·S/(Ks + S) and solving for S gives the steady-state substrate concentration:
The steady-state substrate depends only on the kinetic parameters (μmax, Ks) and the dilution rate, not on the feed concentration Sf. This is a key insight: increasing Sf does not change the residual substrate (until Sf is so low it limits biomass).
The steady-state biomass concentration comes from the substrate mass balance:
Where YX/S is the biomass yield coefficient (g cells per g substrate consumed). At low D, nearly all feed substrate is consumed (S* ≈ 0) and X* ≈ YX/S·Sf. As D approaches μmax, S* rises sharply and X* drops toward zero.
Worked Example: E. coli Chemostat at D = 0.3 h−1
Given: μmax = 0.7 h−1, Ks = 4 mg/L (0.004 g/L), Sf = 10 g/L glucose, YX/S = 0.5 g/g
Step 1. Steady-state substrate:
S* = 0.004 × 0.3 / (0.7 − 0.3)
S* = 0.0012 / 0.4 = 0.003 g/L = 3 mg/L
Step 2. Steady-state biomass:
X* = 0.5 × (10 − 0.003)
X* = 0.5 × 9.997 = 4.999 g/L ≈ 5.0 g/L
Step 3. Volumetric productivity:
D·X* = 0.3 × 5.0 = 1.5 g/L/h
Interpretation: At D = 0.3 h−1, glucose is nearly fully consumed (3 mg/L residual from 10 g/L feed), biomass is high (5.0 g/L), and the vessel produces 1.5 g of cells per litre per hour. The culture is strongly substrate-limited because Ks for glucose in E. coli is only 4 mg/L.
What Happens at Washout?
Washout is the catastrophic loss of culture that occurs when the dilution rate exceeds the organism's maximum specific growth rate. At D > μmax, cells cannot divide fast enough to replace those leaving through the overflow, and the biomass concentration decays exponentially toward zero.
The critical dilution rate Dcrit is defined as the D at which X* = 0. From the steady-state substrate equation, X* = 0 when S* = Sf, which gives:
When Sf >> Ks (the usual case in bioprocessing where feed substrate is grams per litre and Ks is milligrams per litre), Dcrit ≈ μmax. For E. coli with μmax = 0.7 h−1, Ks = 4 mg/L, and Sf = 10 g/L:
Dcrit = 0.7 × 10 / (0.004 + 10) = 0.6997 h−1 ≈ 0.70 h−1
The washout transition is abrupt. At D = 0.65 h−1 (93% of μmax), biomass is still 4.97 g/L. At D = 0.70 h−1, it drops to zero. This cliff-edge behaviour means that operating near Dcrit for maximum productivity carries real risk: any transient drop in μmax (temperature excursion, pH spike, contamination) can trigger washout.
| D (h−1) | D/μmax | S* (mg/L) | X* (g/L) | D·X* (g/L/h) |
|---|---|---|---|---|
| 0.10 | 14% | 0.67 | 5.00 | 0.50 |
| 0.30 | 43% | 3.0 | 5.00 | 1.50 |
| 0.50 | 71% | 10.0 | 4.995 | 2.50 |
| 0.60 | 86% | 24.0 | 4.988 | 2.99 |
| 0.65 | 93% | 52.0 | 4.974 | 3.23 |
| 0.68 | 97% | 136 | 4.932 | 3.35 |
| 0.6997 | ~100% | 10,000 | 0 | 0 |
What Dilution Rate Gives Maximum Productivity?
Maximum volumetric productivity D·X occurs at Dopt, found by differentiating D·X* with respect to D and setting the result to zero. The analytical solution is:
When Sf >> Ks, the square root term is tiny and Dopt approaches μmax. For E. coli with μmax = 0.7, Ks = 0.004, Sf = 10:
Dopt = 0.7 × (1 − √(0.004/10.004))
Dopt = 0.7 × (1 − 0.020) = 0.7 × 0.980 = 0.686 h−1
This is 98% of Dcrit, leaving almost no safety margin. In practice, engineers operate at 60 to 80% of μmax to accommodate disturbances. The productivity penalty for this conservatism is modest: at D = 0.5 h−1 (71% of μmax), productivity is 2.50 g/L/h versus 3.36 g/L/h at Dopt, a 26% reduction that buys a large safety buffer.
For growth-associated products where qp is proportional to μ, the maximum product output rate also occurs near Dopt. For non-growth-associated products (where qp is independent of μ or peaks at low μ), the optimal D shifts much lower, sometimes to 10 to 30% of μmax.
How Long Does It Take a Chemostat to Reach Steady State?
A chemostat requires 4 to 5 residence times to reach steady state after a change in operating conditions. One residence time equals 1/D hours, so higher dilution rates mean faster equilibration.
The transient dynamics after a step change in D follow the differential equation dX/dt = (μ − D)X. Because μ itself depends on S (which is changing), the trajectory is non-linear. At low D the approach is monotonic: biomass slowly rises as substrate is consumed. Near Dcrit, overshoots and oscillatory damped responses can occur because the system's time constants (growth, substrate uptake, mixing) interact.
The practical rule is: do not sample for steady-state data until at least 5 residence times have passed after the last disturbance. Steady state is confirmed when biomass, substrate, and (if measured) product concentrations vary less than 5% over two consecutive residence times. For a glucose-limited E. coli chemostat at D = 0.1 h−1, that means waiting at least 50 hours. At D = 0.5 h−1, only 10 hours.
Turbidostat, Auxostat, and Other Continuous Culture Variants
The chemostat is not the only continuous culture design. Two important variants invert the control logic: instead of fixing D and letting biomass settle, they fix a culture property and let D adjust.
Turbidostat maintains a constant biomass concentration by using an optical density sensor to control the feed pump. When OD rises above the setpoint, the pump speeds up (higher D); when OD falls, the pump slows. At steady state, D settles to a value equal to the organism's growth rate at the set biomass level. Because the culture is not substrate-limited, cells grow near μmax. Turbidostats are preferred for adaptive evolution experiments at maximum growth rate and for organisms that are difficult to grow under nutrient limitation (wall growth, flocculation).
Auxostat is a broader category where any environmental signal (pH, dissolved oxygen, CO2 evolution rate) drives the feed pump. A pH-auxostat increases the feed rate when metabolic acid production drops below a threshold (indicating substrate depletion). A DO-auxostat speeds up the pump when dissolved oxygen rises (less respiratory demand). Auxostats are self-optimizing and can track shifting μmax values during evolution experiments.
| Parameter | Chemostat | Turbidostat | Auxostat |
|---|---|---|---|
| Control variable | D (pump speed fixed) | X (OD setpoint) | pH, DO, or CO2 |
| Adjusted variable | X, S (settle to SS) | D (pump adjusts) | D (pump adjusts) |
| Growth rate | μ = D < μmax | μ ≈ μmax | μ = self-selected |
| Nutrient limitation | Yes (by design) | No (nutrient excess) | Varies |
| Washout risk | If D > μmax | Very low | Low |
| Ideal for | Physiology, kinetics | Evolution at μmax | Tracking μmax shifts |
| Sensor requirement | None (open-loop) | OD probe | pH / DO probe |
| Mathematical model | Monod + mass balance | X-setpoint + PID | Signal-setpoint + PID |
Industrial Applications of Continuous Culture
Continuous culture extends well beyond the laboratory chemostat. Industrial applications span three broad categories: production, research, and waste treatment.
- Single-cell protein (SCP): Quorn (Fusarium venenatum) has been produced in airlift-loop continuous fermenters since the 1980s at volumes exceeding 150 m3, running continuously for over 1,000 hours per campaign.
- Ethanol: Continuous fermentation with cell recycle achieves volumetric productivities of 20 to 50 g/L/h, far exceeding the 2 to 3 g/L/h typical of batch.
- Wastewater treatment: Activated sludge processes are essentially large-scale chemostats where the dilution rate is set by influent flow and the SRT (solids retention time) controls the microbial community.
- Biologics: Perfusion cell culture for monoclonal antibodies operates on chemostat principles with cell retention, achieving 40 to 80 × 106 cells/mL and 1.0 to 2.3 g/L/day volumetric productivity.
- Adaptive evolution: Long-term chemostat cultures (500 to 2,000 generations) select for mutations that improve fitness under the imposed nutrient limitation, generating strains with improved substrate uptake or stress tolerance.
The chemostat concept also underpins modern perfusion bioreactors used in biopharmaceutical manufacturing. The key modification is adding a cell retention device (ATF or TFF), which decouples the cell residence time from the liquid residence time. This creates a "chemostat with cell recycle" where Dliquid can exceed μmax without washout, because cells are retained while spent medium is removed.
Perfusion Calculator
Model perfusion rate, bleed rate, and cell retention for continuous cell culture. The chemostat with cell recycle.
Growth Curve Fitter
Fit μmax and lag time from batch OD data. Use the fitted μmax to set your chemostat's Dcrit and Dopt.
Frequently Asked Questions
What is the dilution rate in a chemostat?
The dilution rate D is the ratio of volumetric flow rate to culture volume: D = F/V, with units of h−1. It represents how many culture volumes pass through the vessel per hour. At steady state in a chemostat, the specific growth rate μ equals D, so by controlling the feed pump you directly set the growth rate. For example, pumping 100 mL/h through a 1 L vessel gives D = 0.1 h−1, fixing the culture's growth rate at 0.1 h−1.
What happens at washout in a chemostat?
Washout occurs when the dilution rate D exceeds the maximum specific growth rate μmax of the organism. Above this critical dilution rate Dcrit (which equals μmax when feed substrate is much greater than Ks), cells are removed from the vessel faster than they can divide. The biomass concentration drops to zero while substrate concentration rises to the feed level Sf. The transition is abrupt: even a small increase beyond Dcrit causes complete culture loss within a few residence times.
How long does it take a chemostat to reach steady state?
A chemostat typically requires 4 to 5 residence times (where one residence time = 1/D) to reach steady state after a change in dilution rate. For E. coli at D = 0.3 h−1, one residence time is 3.3 hours, so steady state is reached in roughly 13 to 17 hours. At lower dilution rates the wait is proportionally longer: D = 0.05 h−1 means 80 to 100 hours. Steady state is confirmed when biomass concentration varies less than 5% over two consecutive residence times.
What is the difference between a chemostat and a turbidostat?
A chemostat fixes the dilution rate (constant feed pump speed) and lets biomass settle to a steady state determined by the limiting nutrient. A turbidostat fixes the biomass concentration by continuously adjusting the dilution rate via an optical density feedback loop. Chemostats operate below μmax and are substrate-limited, making them ideal for studying nutrient-limited physiology. Turbidostats operate near μmax under nutrient excess, making them better suited for adaptive evolution at maximum growth rate.
What dilution rate gives maximum productivity in a chemostat?
Maximum volumetric productivity D·X occurs at Dopt = μmax(1 − √(Ks/(Ks + Sf))), which is typically 80 to 95% of Dcrit. For E. coli with μmax = 0.7 h−1, Ks = 4 mg/L, and Sf = 10 g/L: Dopt ≈ 0.686 h−1. Operating slightly below the washout cliff maximizes output per unit volume per hour, but leaves little margin for disturbances. In practice, engineers set D at 60 to 80% of μmax for safety.
Related Tools
- Perfusion Calculator — Model perfusion rate, bleed rate, and CSPR for continuous cell culture with cell retention.
- Fed-Batch Calculator — Design exponential feeding profiles using μset from chemostat characterization.
- Growth Curve Fitter — Fit μmax from batch OD data to determine your chemostat's Dcrit.
References
- Novick A, Szilard L. Description of the Chemostat. Science. 1950;112(2920):715-716. doi:10.1126/science.112.2920.715
- Hoskisson PA, Hobbs G. Continuous culture — making a comeback? Microbiology. 2005;151(10):3153-3159. doi:10.1099/mic.0.27924-0
- Gresham D, Hong J. The functional basis of adaptive evolution in chemostats. FEMS Microbiology Reviews. 2014;39(1):2-16. doi:10.1111/1574-6976.12082
- Bull AT. The renaissance of continuous culture in the post-genomics age. Journal of Industrial Microbiology and Biotechnology. 2010;37(10):993-1021. doi:10.1007/s10295-010-0816-4