Every gram of product a bioreactor makes is decided by how well the culture is held in its productive window. Cell culture monitoring and control is the discipline of measuring that window and keeping the culture inside it. This guide is the hub for the whole topic: what to measure, how the measurement changes when you move from suspension to adherent to microcarrier systems, and how monitoring becomes closed-loop control. It is the parameter-and-control companion to our deeper dive on cell growth monitoring in suspension culture.
Monitoring vs. control: what each word means
Monitoring is measurement; control is action taken on that measurement. The two are the halves of a feedback loop, and keeping them distinct is the first step to designing a sound strategy. Cell culture monitoring tells you the current viable cell density, dissolved oxygen, pH, and nutrient levels. Control uses those readings to drive an actuator — a gas valve, a base pump, a feed pump — so a parameter stays at its setpoint.
The distinction matters because the two are funded and engineered differently. A parameter can be monitored without being controlled (you watch lactate rise but do not act on it directly), but a parameter can never be controlled without first being monitored. Regulators frame this through Process Analytical Technology (PAT): critical process parameters (CPPs) are monitored, and the ones with the most leverage over critical quality attributes are placed under control. For the full regulatory picture, see our guide to PAT for bioreactor monitoring.
In practice, bioprocess parameters fall into three control tiers. Tier 1 — temperature, DO, pH — is controlled on essentially every bioreactor in the world. Tier 2 — viable cell density, glucose — is increasingly controlled in advanced processes through feedback feeding and perfusion. Tier 3 — lactate, ammonium, osmolality, product titer — is mostly monitored to inform decisions rather than controlled directly, though that frontier is moving fast.
What to monitor: the core parameters
The parameters worth measuring divide into physicochemical conditions, cell-state indicators, and nutrient/metabolite concentrations. The table below is the working reference: what each parameter tells you, its typical setpoint or range in a mammalian process, how it is measured, and how often.
| Parameter | What it tells you | Typical range / setpoint | How measured | Frequency |
|---|---|---|---|---|
| Temperature | Metabolic rate, growth vs. production | 36.5–37 °C (often shifted to 30–33 °C in production) | Pt100 / thermowell | Continuous |
| Dissolved oxygen (DO) | Oxygen sufficiency vs. limitation | 30–50% air saturation | Optical or polarographic probe | Continuous |
| pH | Acid/base balance, CO2 & lactate load | 6.8–7.2 | Glass or optical pH probe | Continuous |
| Viable cell density (VCD) | Growth, the timing of feed and harvest | 0.3–20×106 cells/mL | Counter (offline) or capacitance (in-line) | 1–2×/day or continuous |
| Viability | Onset of death phase | >90% until late culture | Trypan-blue exclusion / image cytometry | 1–2×/day |
| Glucose | Carbon sufficiency; drives feeding | Hold 2–6 g/L | Bioanalyser (at-line) or Raman (in-line) | 1–3×/day or continuous |
| Lactate | Metabolic state; the metabolic shift | <2 g/L desirable | Bioanalyser or Raman | 1–3×/day or continuous |
| Ammonium | Glutamine metabolism, toxicity | <3–5 mM | Bioanalyser | 1×/day |
| Osmolality | Cumulative feed/base load on cells | 280–400 mOsm/kg | Osmometer (offline) | 1×/day |
| Dissolved CO2 | Stripping efficiency, pH coupling | 30–100 mmHg | Severinghaus or optical probe | Continuous / at-line |
Two relationships in that table drive most real decisions. First, glucose and lactate together define the metabolic state: a healthy CHO culture often shifts from lactate production to lactate consumption mid-run, and watching that shift is more informative than either value alone — the topic of our deep dive on lactate accumulation in CHO. Second, DO and pH control loops interact: sparging to hold DO strips CO2, which raises pH, so the two loops must be tuned together. Our article on dissolved oxygen in CHO culture covers that coupling in detail.
Figure 2. A typical 14-day fed-batch profile: VCD peaks near day 10, then viability falls as the culture enters death phase. High-frequency monitoring is what flags the inflection early.
Monitoring by culture format: suspension, adherent, microcarrier
The single biggest variable in a monitoring strategy is whether the cells are in suspension, attached to a surface, or growing on microcarriers. The physicochemical parameters are measured the same way in all three, but cell-state measurement changes completely. This is the question behind searches like "monitoring and control of adherent cell culture" or "microcarrier cell culture monitoring," and the answer is format-specific.
Suspension culture is the easy case. The vessel is homogeneous, so a single sample represents the whole culture and VCD is read directly by a counter or in-line capacitance probe. This is why stirred-tank suspension processes dominate monoclonal antibody manufacturing. Most of the in-line toolkit — capacitance, Raman, optical density — was developed for and works best in suspension.
Adherent culture is the hard case for monitoring. Cells are attached to flasks, roller bottles, or a fixed bed, so you cannot sample them for a count without trypsinising. Growth is instead tracked by microscopy and confluence imaging, or inferred indirectly from the medium: glucose consumption rate and lactate production rate are proxies for biomass when a direct count is impossible. Recent work has even put fully integrated wireless multivariate sensors into adherent stem-cell systems for long-term in-situ monitoring without disturbing the monolayer.
Microcarrier culture sits between the two. Because the microcarrier beads are suspended in a well-mixed slurry, bulk DO, pH, and metabolite probes behave just like suspension — a representative sample is straightforward. But the cells are attached to beads, so cell number requires either detaching and counting nuclei or imaging the beads in situ. In-situ microscopy now lets process developers visualise confluence on microcarriers in real time, which is a meaningful advance for vaccine and cell-therapy processes (see our microcarrier cell culture guide).
Not sure which sensor fits your process?
Answer six questions about scale, organism, and culture format and get a ranked sensor recommendation from 11 probe types.
Offline, at-line, in-line: the measurement spectrum
Every monitoring method sits somewhere on a trade-off between how much it tells you and how fast it tells you. There are three positions on that spectrum, and a good strategy uses all three rather than betting on one.
- Offline — a sample is taken to a separate instrument (osmometer, full metabolite panel, image cytometer). Richest information, but a latency of minutes to hours and a labour cost per sample. This is where osmolality, ammonium, and reference VCD live.
- At-line — measured next to the bioreactor on a benchtop analyser within a few minutes, often via an autosampler. The classic example is a bioanalyser reporting glucose, lactate, and gases. See in-line vs at-line glucose monitoring for the practical comparison.
- In-line / in-situ — a probe in the vessel reports continuously with seconds of latency and no sampling. DO, pH, temperature, and capacitance biomass are in-line by default; Raman brings glucose, lactate, VCD, and titer in-line through a chemometric model. This is the only class fast enough to close a real-time control loop.
The reason in-line monitoring matters for cell culture monitoring and control is latency: you cannot control what you measure once a day. A capacitance probe that reports viable biomass every few seconds can drive a perfusion bleed or trigger a feed; a daily offline count cannot. The spectroscopic in-line tools are covered in our articles on Raman spectroscopy for bioprocess monitoring and biomass sensors for bioreactors.
From monitoring to closed-loop control
Closed-loop control is monitoring plus an automatic correction: a sensor measures, a controller compares the reading to a setpoint, and an actuator drives the deviation back to zero. The same loop in Figure 1 is reused for every controlled parameter — only the sensor and actuator change.
The two universal loops are the DO cascade and pH control. When DO falls below setpoint, the controller first raises agitation, then increases the oxygen fraction or sparge rate — a cascade because it escalates through cheaper actions before expensive ones. When pH drifts acidic (lactate, CO2), base is metered in; when it drifts basic, CO2 is sparged. These two are tuned together because, as noted above, oxygen sparging strips CO2 and nudges pH. Our bioreactor instrumentation guide details the probes and calibration behind these loops.
The advanced layer is biomass- and nutrient-based control. A capacitance probe reading viable cell volume can drive an automated cell bleed to hold VCD constant in perfusion, a strategy validated decades ago with radio-frequency impedance and now routine. Glucose can be held at setpoint by feedback feeding, where an at-line or in-line glucose reading triggers a calculated bolus or adjusts a continuous feed rate — turning the classic open-loop fed-batch into a closed loop.
Worked Example: Glucose Feedback Feeding
A 5.0 L CHO fed-batch reads 2.0 g/L glucose at the morning sample. The target is to restore 4.0 g/L using a 400 g/L glucose feed stock. How much feed?
feed volume = 10 g ÷ 400 g/L = 0.025 L = 25 mL
dilution check = 25 mL / 5000 mL = 0.5% (negligible)
A controller doing this automatically would meter 25 mL on each reading, or convert the deficit into a continuous feed rate. This is exactly the feedback logic that distinguishes a controlled fed-batch from a fixed-bolus one. Plan and size feeds with the Fed-Batch Calculator.
Figure 4. Glucose held in range by feeding while lactate rises, plateaus at the metabolic shift, then re-accumulates as the culture declines. Controlling glucose tightly is one lever on the lactate trajectory.
Building a monitoring and control strategy
A workable strategy layers the three measurement classes against the parameters that matter most for your product. Start from control needs and work outward: anything you intend to control must be measured in-line; anything that informs a daily decision can be at-line; anything that is a release or characterisation parameter can be offline.
- Lock the Tier 1 loops first. Temperature, DO, and pH probes plus their actuators are non-negotiable. Calibrate them every batch (or use pre-calibrated single-use sensors).
- Add a viable-biomass signal. An in-line capacitance probe gives continuous VCD for control; pair it with a once-daily offline count as the reference. Choose between probe types with our comparison of capacitance vs optical biomass sensors.
- Add nutrient visibility. At-line glucose/lactate at minimum; in-line Raman if the process justifies the model-building effort.
- Decide what to control vs. monitor. Glucose feedback feeding and capacitance-driven bleed are the highest-value control additions; lactate and osmolality usually stay monitored.
- Use soft sensors to fill gaps. Where a hardware probe is impractical, a soft sensor can estimate a hard-to-measure variable (like VCD or titer) from the parameters you already measure.
Match the method to the scale, too. A 50 mL ambr microbioreactor cannot give up 1 mL/day for offline counts the way a 2,000 L vessel can, so small-scale work leans harder on in-line and image-based methods. The format and scale together set the realistic monitoring envelope.
Turn monitoring data into a growth rate
Paste VCD readings and the Growth Curve Fitter returns specific growth rate, doubling time, and the phase boundaries that trigger your control actions.
Where cell culture monitoring is heading
The direction of travel is from single-point probes toward model-rich, real-time, multivariate monitoring. Three threads stand out as of 2026.
Spectroscopy plus chemometrics is maturing from glucose/lactate into simultaneous prediction of VCD, viability, titer, and even product quality attributes from one Raman probe, replacing a rack of single-analyte instruments. Soft sensors and digital twins combine the cheap measurements you already have with a process model to estimate the expensive ones and to predict where the culture is heading, not just where it is. And novel non-invasive sensing — wireless multivariate sensor arrays, magnetic resonance relaxometry, Doppler-ultrasound cell counting — is extending real-time monitoring into formats (adherent, sealed, single-use) where probes were never practical.
The common thread is closing the loop faster and on more variables. As more parameters move from offline to in-line, more of them become controllable — and cell culture monitoring and control converge into a single real-time activity.
Frequently Asked Questions
What is the difference between cell culture monitoring and control?
Monitoring is measuring the state of a culture (VCD, DO, pH, glucose, lactate). Control is acting on those measurements to hold a parameter at a setpoint — sparging oxygen when DO falls, or feeding when glucose drops. Monitoring is the sensor half of the loop; control is the actuator half. A parameter can be monitored without being controlled, but not controlled without first being monitored.
What parameters are monitored in cell culture?
The core controlled parameters are temperature, dissolved oxygen, and pH, held at setpoint on essentially every bioreactor. The key state indicators monitored but not always controlled are viable cell density, viability, glucose, lactate, ammonium, osmolality, and dissolved CO2. Temperature, DO, and pH are continuous; VCD and metabolites are continuous (in-line) or once or twice daily (sampling).
How do you monitor adherent and microcarrier cultures differently from suspension?
Suspension gives a representative sample, so VCD is read directly. Adherent culture cannot be sampled for a count without detachment, so growth is tracked by microscopy, confluence imaging, or indirect proxies (glucose consumption, lactate production). Microcarrier culture allows bulk DO/pH/metabolite probes because the bead slurry is well mixed, but cell number needs nuclei counting after detachment or in-situ imaging of the beads.
How often should viable cell density be measured?
Offline VCD by trypan-blue exclusion or an automated counter is typically measured once a day in development and once or twice daily in manufacturing, timed to inform feeding and harvest. In-line capacitance probes measure viable biomass continuously, every few seconds, which catches a stalling culture hours before the next scheduled sample would.
What is closed-loop control in a bioreactor?
Closed-loop control is an automatic feedback loop: a sensor measures, a controller compares to setpoint, and an actuator corrects the deviation. Classic examples are the DO cascade (raise agitation, then oxygen) and pH control (add base or sparge CO2). Advanced loops use capacitance to control cell bleed in perfusion or glucose feedback to drive feeding.
Can cell culture be monitored in real time?
Yes. Temperature, DO, and pH are real-time on every bioreactor. Viable biomass is real-time via capacitance, and glucose, lactate, VCD, and titer can be measured in-line and in real time by Raman spectroscopy with a calibrated model. Real-time in-line monitoring is the basis of the PAT framework for advanced process control.
Related Tools
References
- Fung Shek C. & Betenbaugh M. (2021). Taking the pulse of bioprocesses: at-line and in-line monitoring of mammalian cell cultures. Current Opinion in Biotechnology, 71, 191–197. doi:10.1016/j.copbio.2021.08.007
- Rathore A.S., Mishra S., Nikita S., et al. (2021). Bioprocess Control: Current Progress and Future Perspectives. Life, 11(6), 557. doi:10.3390/life11060557
- Carvell J.P. & Dowd J.E. (2006). On-line Measurements and Control of Viable Cell Density in Cell Culture Manufacturing Processes using Radio-frequency Impedance. Cytotechnology, 50, 35–48. doi:10.1007/s10616-005-3974-x
- Lomont J.P., Love E., Wagner L., Ralbovsky N.M., et al. (2026). In Situ Microscopy for Real-Time Visualization of Microcarrier Cell Cultures for Live Virus Vaccine Process Development. Biotechnology and Bioengineering, 123(4), 950–960. doi:10.1002/bit.70127
- Lee S., Kim S., Lim H., Kim Y., et al. (2024). Large-scale smart bioreactor with fully integrated wireless multivariate sensors and electronics for long-term in situ monitoring of stem cell culture. Science Advances, 10(7), eadk6714. doi:10.1126/sciadv.adk6714