Every critical process parameter in a bioreactor depends on a sensor reading it correctly. A DO probe that drifts 5% undetected shifts your oxygen setpoint enough to reduce CHO cell viability. A pH electrode fouled by protein can mask a 0.2 pH unit excursion that alters glycosylation profiles. Bioreactor instrumentation is the foundation of process control, yet it receives less attention than the biology it serves.
This guide covers the six sensor types installed on a typical production bioreactor: dissolved oxygen, pH, temperature, biomass, pressure, and off-gas composition. For each, you will find how the measurement principle works, which sensor technologies are available, how to calibrate correctly, what drift rates to expect, and when to replace the probe.
The Six Core Bioreactor Sensors
A fully instrumented stirred-tank bioreactor carries sensors for dissolved oxygen (DO), pH, temperature, viable cell density (biomass), headspace pressure, and exhaust gas composition. Together, these six measurements provide enough real-time data to run closed-loop control of all critical process parameters in upstream bioprocessing.
The table below summarizes each sensor type, its measurement principle, typical accuracy, and the parameter it controls.
| Sensor | Parameter | Principle | Typical Accuracy | Response Time |
|---|---|---|---|---|
| DO (optical) | Dissolved O2 | Fluorescence quenching | ± 1% air sat. or ± 0.05 mg/L | < 30 s (t90) |
| DO (polarographic) | Dissolved O2 | Amperometric (Clark cell) | ± 1-2% air sat. | 30-90 s (t90) |
| pH (glass) | pH | Potentiometric (Nernst) | ± 0.02 pH (fresh), ± 0.1 pH (end of batch) | < 15 s |
| Temperature (RTD) | Temperature | Resistance vs. temperature (Pt100) | ± 0.1 °C (Class A), ± 0.3 °C (Class B) | 5-30 s (in thermowell) |
| Biomass (capacitance) | Viable cell density | Dielectric spectroscopy | ± 5-10% of reading | < 60 s |
| Biomass (turbidity) | Total cell density / OD | NIR backscatter or transmission | ± 5-15% of reading | < 10 s |
| Pressure | Headspace pressure | Piezoresistive / capacitive | ± 0.1-0.5% full scale | < 1 s |
| Off-gas (paramagnetic + IR) | O2, CO2 in exhaust | Paramagnetic O2, NDIR CO2 | ± 0.01-0.1% vol | < 5 s |
Dissolved Oxygen (DO) Sensors
Dissolved oxygen is the most tightly controlled parameter in aerobic bioprocesses, and DO sensor accuracy directly determines whether your culture receives the oxygen it needs. Two sensor technologies dominate: electrochemical (polarographic/galvanic) and optical (fluorescence quenching).
Electrochemical DO sensors
Polarographic (Clark-type) sensors apply a polarization voltage across a cathode and anode separated from the culture by an oxygen-permeable membrane. Oxygen diffusing through the membrane is reduced at the cathode, generating a current proportional to the dissolved oxygen partial pressure. Galvanic sensors use a dissimilar-metal electrode pair that generates current without an external voltage.
Electrochemical sensors require regular maintenance. The membrane must be replaced every 4-8 weeks depending on medium composition, and the internal electrolyte (typically KCl solution) must be refilled at each membrane change. Signal drift of 2-5% per week is typical because the membrane gradually fouls with proteins and lipids.
Optical DO sensors
Optical sensors measure dissolved oxygen by fluorescence quenching. A luminescent dye (typically a ruthenium or platinum porphyrin complex) is excited by a pulse of light. In the presence of oxygen, the fluorescence lifetime shortens proportionally. The sensor measures this lifetime and converts it to a DO value using the Stern-Volmer relationship.
Optical sensors have largely replaced electrochemical probes in new bioreactor installations. They consume no oxygen during measurement, require no electrolyte, and their drift is typically less than 1% per week. The sensing cap (luminescent dye layer) lasts 1-2 years before the dye degrades enough to require replacement.
DO calibration protocol
Two-point calibration is the standard for DO sensors in bioreactors:
- 100% saturation point: Sparge air through the medium at the process temperature and agitation rate until equilibrium is reached (typically 15-30 minutes). Set this as 100% air saturation.
- Zero point: Sparge high-purity nitrogen (grade 5.0, 99.999%) through the medium until DO reads stable at baseline (typically 5-10 minutes). Alternatively, add excess sodium sulfite (1-2 g/L) to chemically consume all dissolved oxygen.
Calibrate DO probes after autoclaving and at process conditions (temperature, agitation, medium) because oxygen solubility depends on temperature, salinity, and pressure. A probe calibrated in water at 20 °C will read incorrectly in culture medium at 37 °C.
Worked Example: DO Calibration Check
A CHO cell culture runs at 37 °C in DMEM/F12 medium. After 10 days, you suspect DO probe drift.
Verification method: Take an offline sample, measure DO with a calibrated benchtop meter. The benchtop reads 42% air saturation; the inline probe reads 38%.
Drift magnitude: (42 - 38) / 42 = 9.5% relative error.
Action: At > 5% relative deviation, apply a one-point offset correction using the benchtop value as reference. If this correction exceeds the manufacturer's specified range (typically ± 10-15% offset), replace the probe or sensing cap. Log the correction in the batch record for GMP traceability.
pH Sensors
Glass electrode pH sensors are the workhorse of bioreactor pH measurement, installed on virtually every production-scale vessel. The glass membrane develops a voltage proportional to the hydrogen ion activity difference between the process fluid and the internal buffer (Nernst equation). This potentiometric signal is read by a high-impedance transmitter.
pH probe construction for bioreactors
Bioreactor pH probes differ from laboratory bench electrodes in three ways: they are designed to withstand autoclave temperatures (121 °C, 20-30 minutes), they use reinforced glass membranes to resist mechanical stress from impellers and sparging, and they have pressurized or gel-filled reference systems to prevent process fluid from back-diffusing into the reference junction.
Gel-filled reference electrodes are lower-maintenance but drift faster during long runs (above 10 days) because the gel junction clogs with proteins. Pressurized liquid-electrolyte designs maintain a positive outflow of KCl, keeping the junction clear, but require periodic electrolyte refilling.
pH calibration protocol
Calibrate pH probes before every batch using a two-point buffer calibration:
- Buffer 1: pH 7.00 (sets the zero-point offset)
- Buffer 2: pH 4.01 or pH 9.21 (sets the slope)
A healthy probe has a slope of 95-102% of the theoretical Nernst slope (59.16 mV/pH at 25 °C). A slope below 90% or asymmetry potential exceeding ± 30 mV indicates the probe should be replaced. Always calibrate at room temperature and apply automatic temperature compensation (ATC) for the process temperature.
pH drift during fermentation
Protein fouling is the primary cause of pH drift. The glass membrane surface accumulates a protein film that reduces ion exchange efficiency, shifting readings toward pH 7 (the isopotential point). In CHO fed-batch cultures running 14 days, cumulative drift of 0.1-0.3 pH units is common. This drift is insidious because the controller compensates by adding more base, which raises osmolality and can reduce cell viability.
| Symptom | Likely Cause | Corrective Action |
|---|---|---|
| Slow response (> 30 s to stable reading) | Protein fouling on glass membrane | Soak in pepsin/HCl cleaning solution for 1-2 hours |
| Slope < 90% of theoretical | Aged glass membrane or depleted reference | Replace probe (if > 20 autoclave cycles, membrane is degraded) |
| Noisy signal (> 0.05 pH fluctuations) | Reference junction clogging or air bubbles | Refill electrolyte (liquid-fill type); reposition probe away from sparger |
| Reading drifts toward pH 7.0 | Protein film on glass + asymmetry shift | Mid-batch offset correction using offline benchtop measurement |
| Calibration fails (cannot reach both buffers) | Cracked glass membrane or dead probe | Replace probe immediately |
Temperature Sensors
Pt100 resistance temperature detectors (RTDs) are the standard for bioreactor temperature measurement. A Pt100 element has a resistance of 100.00 ohms at 0 °C and changes resistance linearly with temperature (0.385 ohms per °C). This predictable, stable behavior makes RTDs the most accurate temperature sensor for the 15-45 °C range used in bioprocessing.
RTDs are installed in a thermowell welded to the bioreactor shell. The thermowell protects the sensor from direct contact with the culture, allows replacement without breaking sterility, and withstands CIP/SIP conditions. The thermal mass of the thermowell adds 5-30 seconds of lag to the measurement, which matters for fast temperature-shift protocols (such as 37 °C to 32 °C in CHO production phase).
RTD accuracy classes and calibration
IEC 60751 defines two accuracy classes relevant to bioprocessing:
- Class A: ± 0.15 °C at 0 °C (± 0.35 °C at 100 °C). Standard for GMP bioreactors.
- Class B: ± 0.30 °C at 0 °C (± 0.80 °C at 100 °C). Acceptable for R&D bioreactors.
RTDs are the most stable bioreactor sensors. Annual drift is typically less than 0.02 °C for a well-made Pt100, so calibration verification every 6-12 months is sufficient. Calibrate against a traceable reference thermometer (ITS-90) using an ice-point check (0 °C) and a warm-water bath at the process temperature.
Biomass Sensors
Inline biomass sensors enable real-time monitoring of cell concentration without manual sampling. Two technologies are used: capacitance probes (dielectric spectroscopy) for viable cell density, and turbidity probes (optical density) for total cell density.
Capacitance (dielectric spectroscopy) probes
Capacitance probes apply a radio-frequency alternating electric field (typically 0.1-10 MHz) between two electrodes. Intact cell membranes, which are lipid bilayers approximately 7 nm thick, act as tiny capacitors. They polarize in the field and contribute a measurable capacitance signal proportional to the total volume of viable, membrane-enclosed cells. Dead cells with compromised membranes do not polarize and are invisible to the measurement.
This selectivity for viable cells is the key advantage over turbidity, which cannot distinguish live from dead cells or cell debris. Capacitance probes correlate linearly with viable cell density from approximately 0.5 to 50 x 106 cells/mL in mammalian culture (higher ranges for microbial fermentation where cells are smaller). The measurement is expressed in pF/cm (picofarads per centimeter of electrode gap).
Turbidity (optical density) probes
Turbidity probes measure light scattering or absorbance in the near-infrared range (typically 840-910 nm). They report a signal proportional to total suspended solids, including live cells, dead cells, cell debris, and microcarriers. Two configurations exist: transmission (light passes through a fixed optical path) and backscatter (reflected light is measured at an angle).
Turbidity probes are simpler and less expensive than capacitance probes. They work well for microbial fermentation where viability is typically above 95% and total cell density is the relevant metric. For mammalian cell culture, where viability can drop to 60-70% late in fed-batch, turbidity over-reports the viable cell count.
Biomass probe calibration
Both probe types require a cell-count calibration curve built from offline reference measurements (hemocytometer, trypan blue exclusion, or automated cell counter). Take 8-12 paired samples across the full growth range and fit a linear regression of probe signal (pF/cm or AU) versus offline cell count. The calibration is cell-line-specific and must be rebuilt when switching to a different host cell.
Pressure and Off-Gas Sensors
Headspace pressure in a bioreactor is typically maintained at 0.2-0.5 bar gauge to prevent contamination ingress and to improve gas-liquid mass transfer. Pressure sensors (piezoresistive or capacitive transducers) are mounted on the headplate and provide the signal for the backpressure control valve.
Pressure measurement also enables differential-pressure-based level sensing. Two connected pressure sensors, one at the vessel bottom and one in the headspace, measure the differential pressure caused by the liquid column. This provides continuous, non-invasive volume measurement with accuracy of ± 1-2% of the working volume.
Off-gas analysis
Exhaust gas analyzers measure the O2 and CO2 concentrations in the bioreactor off-gas. The oxygen uptake rate (OUR) and carbon dioxide evolution rate (CER) are calculated from the difference between inlet and outlet gas compositions, multiplied by the gas flow rate:
OUR = Fgas × (yO2,in - yO2,out) / VL
CER = Fgas × (yCO2,out - yCO2,in) / VL
RQ = CER / OUR
The respiratory quotient (RQ) is a dimensionless ratio that reveals metabolic state. For mammalian cells on glucose, RQ is typically 0.9-1.0. For E. coli in glucose-limited fed-batch, RQ near 1.0 indicates balanced metabolism, while RQ above 1.0 signals overflow metabolism and acetate production. Off-gas analysis is the only non-invasive way to measure these metabolic rates in real time.
Two analyzer technologies dominate:
- Paramagnetic oxygen analyzers: Exploit the paramagnetic property of O2 molecules. Accuracy ± 0.01-0.02% vol. No consumables.
- NDIR (non-dispersive infrared) CO2 analyzers: CO2 absorbs IR light at 4.26 μm. Accuracy ± 0.01-0.1% vol. Lamp replacement every 2-5 years.
Multiplexed mass spectrometers (quadrupole MS) can monitor 4-16 bioreactors from a single instrument by switching between off-gas lines. They measure O2, CO2, N2, Ar, and volatile metabolites simultaneously but cost 50,000-150,000 USD versus 10,000-30,000 USD for a paramagnetic/NDIR pair.
Single-Use Bioreactor Sensors
Single-use bioreactors (SUBs) use pre-sterilized, disposable bag assemblies that eliminate the need for CIP/SIP. Sensors in SUBs must be integrated into the bag film during manufacturing and must survive gamma irradiation (25-50 kGy). This constraint has driven the development of optical patch sensors for pH and DO.
Optical patch sensors
A fluorescent indicator dye is embedded in a thin polymer patch (5-10 mm diameter) bonded to the inner surface of the bag. An external optical reader, mounted on the outside of the bag wall, excites the dye through the transparent film and measures the fluorescence response. For DO, the dye is a platinum porphyrin (fluorescence lifetime quenched by oxygen). For pH, the dye is a dual-lifetime fluorophore responsive to hydrogen ion concentration.
Optical patches are factory-calibrated. The calibration parameters are printed on the bag label or encoded in an RFID chip. No pre-batch calibration is required. Accuracy after gamma irradiation is typically ± 0.05 pH units and ± 2% DO of reading, with a shelf life of 2-3 years.
| Feature | Reusable Probes | Single-Use Patches |
|---|---|---|
| Sterilization | Autoclave (121 °C, 20-30 min) | Gamma irradiation (25-50 kGy) |
| Calibration | Required before each batch | Factory pre-calibrated |
| Lifespan | 10-30 autoclave cycles (pH); 1-2 years (optical DO) | Single use (disposed with bag) |
| Accuracy (pH) | ± 0.02 pH (fresh) | ± 0.05 pH |
| Accuracy (DO) | ± 1% air sat. | ± 2% of reading |
| Cost per batch | Low (amortized over cycles) | 20-100 USD per sensor patch |
| Maintenance | Cleaning, electrolyte refill, membrane change | None |
| Drift (14-day run) | 0.1-0.3 pH; 2-5% DO | < 0.05 pH; < 2% DO |
Calibration Schedule and Maintenance Intervals
A consistent calibration and maintenance schedule is the single most effective way to prevent sensor-related batch deviations. The chart below shows typical drift rates for each sensor type over a 14-day CHO fed-batch run, illustrating why some sensors need daily verification while others are stable for months.
| Sensor | Calibration Frequency | Maintenance Task | Replacement Interval |
|---|---|---|---|
| DO (optical) | Before each batch; verify every 1-3 batches | Inspect sensing cap for discoloration | Sensing cap every 1-2 years; body lasts 5-10 years |
| DO (polarographic) | Before each batch (two-point) | Replace membrane every 4-8 weeks; refill electrolyte | Cathode repolishing every 6-12 months; replace probe after 3-5 years |
| pH (glass) | Before each batch (two-point buffer) | Clean with pepsin/HCl after each batch; store in pH 4 or KCl | Every 10-30 autoclave cycles (typically 6-18 months) |
| Temperature (Pt100 RTD) | Verify every 6-12 months (ice-point check) | Inspect thermowell for corrosion | 5-10 years (very stable) |
| Biomass (capacitance) | Rebuild calibration curve per cell line | Clean electrodes with 70% ethanol between batches | Probe body 3-5 years; electrode coating as needed |
| Pressure | Annual verification against reference gauge | Check diaphragm seal integrity | 5-10 years |
| Off-gas (paramagnetic + NDIR) | Monthly span check with certified gas standard | Replace sample line filters; check for condensation | IR lamp every 2-5 years; paramagnetic cell 5-10 years |
Worked Example: Pre-Batch Sensor Qualification Checklist
Before starting a GMP CHO fed-batch run in a 200 L stainless-steel bioreactor, verify each sensor:
- pH probe: Two-point calibration (pH 7.00 and 4.01). Slope = 97.2% (acceptable: > 90%). Asymmetry = +8 mV (acceptable: < ± 30 mV). Record in batch record. ✓ PASS
- DO probe (optical): Two-point calibration. 100% point set with air sparge at 37 °C, 150 rpm, in medium. Zero point set with N2 sparge. Response time t90 = 18 s (spec: < 30 s). ✓ PASS
- Temperature (Pt100): Ice-point check: reads 0.05 °C (acceptable: ± 0.15 °C Class A). Last full calibration: 4 months ago. ✓ PASS
- Biomass (capacitance): Zeroed in sterile medium (0 pF/cm baseline). Calibration curve for CHO-K1 loaded from previous qualification (R² = 0.993). ✓ PASS
- Pressure: Reads 0.00 bar gauge at atmospheric. Backpressure valve cycles correctly to 0.3 bar setpoint. ✓ PASS
- Off-gas: Span-checked with 5.00% CO2 / 15.00% O2 / balance N2 certified standard. Reads 5.01% CO2, 14.98% O2. Within ± 0.1% spec. ✓ PASS
Total qualification time: approximately 90 minutes. All six sensors pass. The batch may proceed.
OTR/kLa Estimator
Calculate oxygen transfer rates from your DO sensor data. Estimate kLa using dynamic gassing-out or steady-state methods.
Gas Mixing Calculator
Size gas blending for O2 enrichment and CO2 overlay. Match your off-gas analyzer readings to inlet composition.
Scale-Up Calculator
Scale bioreactor parameters from bench to production. Match P/V, tip speed, kLa, and mixing time across vessel sizes.
Related Tools
- Bioreactor Data Dashboard — Visualize and overlay process data from multiple bioreactor runs.
- Growth Curve Fitter — Fit growth curves and extract specific growth rate from biomass sensor data.
- Heat Transfer Calculator — Size jackets and coils using temperature sensor data for thermal control.
References
- Cui G., Wu R., Cao L., Abedin S., Goel K., Yoon S. & Wang X. (2025). Optical Fiber pH and Dissolved Oxygen Sensors for Bioreactor Monitoring: A Review. Sensors, 26(1), 10. doi:10.3390/s26010010
- Busse C., Biechele P., de Vries I., Reardon K.F., Solle D. & Scheper T. (2017). Sensors for disposable bioreactors. Engineering in Life Sciences, 17(8), 940-952. doi:10.1002/elsc.201700049
- O'Mara P., Farrell A., Bones J. & Twomey K. (2018). Staying alive! Sensors used for monitoring cell health in bioreactors. Talanta, 176, 130-139. doi:10.1016/j.talanta.2017.07.088
- Metze S., Ruhl S., Greller G., Grimm C. & Scholz J. (2020). Monitoring online biomass with a capacitance sensor during scale-up of industrially relevant CHO cell culture fed-batch processes in single-use bioreactors. Bioprocess and Biosystems Engineering, 43, 193-205. doi:10.1007/s00449-019-02216-4
- Saxena N., Mishra S., Gupta K., Runkana V., Gomes J. & Rathore A.S. (2023). Advances in bioreactor control for production of biotherapeutic products. Biotechnology and Bioengineering, 120(5), 1189-1214. doi:10.1002/bit.28346
Frequently Asked Questions
How often should I calibrate DO probes in a bioreactor?
For polarographic DO sensors, calibrate before every batch using a two-point method: 100% air saturation and zero (nitrogen sparge or sodium sulfite solution). Optical DO sensors are more stable and typically need calibration verification every 1-3 batches, with full recalibration only when drift exceeds 2% of the reading. Always calibrate after membrane or electrolyte replacement on electrochemical sensors.
What causes pH probe drift during fermentation?
The main causes are protein fouling of the glass membrane and reference junction clogging, which restrict ion exchange and shift the measured potential. High-cell-density cultures (above 50 x 106 cells/mL) accelerate fouling. Autoclave cycling also degrades the glass membrane over time, typically limiting probe lifespan to 10-30 sterilization cycles. Gel-filled reference electrodes drift faster than pressurized liquid-electrolyte designs in long-duration processes.
Should I use optical or electrochemical DO sensors for my bioreactor?
Optical (fluorescence quenching) DO sensors are preferred for most modern bioreactor applications. They have no electrolyte to consume, no membrane to replace, drift less than 1% per week versus 2-5% for polarographic sensors, and respond in under 30 seconds. Electrochemical sensors remain useful when budget is constrained or when measuring very low DO (below 0.1% air saturation) where some optical sensors lose resolution.
How do capacitance biomass probes work?
Capacitance probes apply a radio-frequency electric field (typically 0.1-10 MHz) across two electrodes immersed in the culture. Intact cell membranes act as tiny capacitors, polarizing in the field. The measured capacitance is directly proportional to the total volume of viable, membrane-bound cells. Dead cells with compromised membranes do not contribute, making capacitance probes selective for viable cell density.
Can bioreactor sensors be used in single-use bioreactors?
Yes. Single-use bioreactors use pre-calibrated, gamma-irradiated optical patch sensors for pH and DO that are integrated into the bag film during manufacturing. These patches are read through the transparent bag wall by an external optical reader. They eliminate the need for autoclavable probes and pre-batch calibration. Shelf life is typically 2-3 years after gamma irradiation, and accuracy is within ± 0.05 pH units and ± 2% DO of the reading.