Engineering Guide · Vendor-Neutral

Severinghaus vs Optical CO2 Sensors: Which to Use in Bioreactors

Severinghaus electrochemical CO2 sensor versus optical CO2 sensor — side-by-side mechanism diagram pH electrode HCO₃⁻ electrolyte Gas-permeable membrane CO₂ CO₂ + H₂O → pH shift Severinghaus electrode Potentiometric · electrolyte + membrane Industry standard since 1957 VS IR LED + PD IR sensing layer (or HPTS dye) CO₂ CO₂ absorbs IR (or quenches dye) Optical CO₂ sensor Solid-state · no consumables Single-use ready · fluorescence or mid-IR
Figure 1: Severinghaus electrodes (left) measure CO2 indirectly — gas diffuses through a permeable membrane into a bicarbonate electrolyte, where the resulting pH shift is read by an internal pH glass electrode. Optical CO2 sensors (right) detect CO2 either directly via mid-IR absorption (Hamilton CO2NTROL) or indirectly via fluorescence quenching of an HPTS dye in a buffer (PreSens) — both eliminate the consumable electrolyte.
Quick Verdict

Severinghaus electrodes remain the installed-base default; optical CO2 sensors are the safer pick for new builds. The Mettler InPro 5000i Severinghaus probe owns the cGMP installed base and the wider 0–1000 mbar range microbial fermentation needs. Solid-state mid-IR optical (Hamilton CO2NTROL) eliminates electrolyte refills and membrane changes for stainless steel mAb fed-batch, and HPTS-dye optical (PreSens) is the only realistic option in single-use bags. The DO version of this comparison goes the same way — see Optical vs Polarographic DO.

Key differences at a glance

Side-by-side comparison

FactorSeveringhaus electrodeOptical CO2 sensor
Measurement principlePotentiometric — pH shift in bicarbonate electrolyteMid-IR absorption (solid-state) or fluorescence quenching of HPTS dye
Internal consumablesBicarbonate electrolyte + gas-permeable membraneNone (solid-state) or factory-sealed dye
Maintenance intervalElectrolyte every 3–6 months, membrane every 6–12 monthsMaintenance-free for probe lifetime
Typical accuracy±10% (10–900 mbar)±0.5–1.2 mmHg at physiological pCO2 (HPTS)
Working range0–1000 mbar (0–760 mmHg)10–250 hPa (HPTS) / 5–1000 mbar (mid-IR)
Response time (t90)60–120 s at 25°C60–180 s at 25°C
SterilisationAutoclave + SIP up to 130°CSIP to 140°C (mid-IR) / gamma (single-use spots)
Single-use bag compatibilityNot availableSP-CD1 spots + FTC-SU-CD1 flow cells
Calibration drift between cyclesModerate — electrolyte dilution and membrane foulingLow — solid-state has no internal phase to drift
Typical capital cost (per channel)£3,000–£5,000£4,000–£7,500

Values reflect typical published specifications. Vendor datasheets take precedence for specific instrument specs.

How Severinghaus CO2 electrodes work

The Severinghaus electrode was introduced in 1957 by Dr. John W. Severinghaus and has been the industry-standard dissolved CO2 sensor in bioprocess for more than five decades. The basic principle is indirect: CO2 cannot be measured potentiometrically on its own, but its hydration product carbonic acid can be measured through pH. A thin film of bicarbonate electrolyte is held behind a gas-permeable membrane (typically silicone-reinforced PTFE). CO2 from the culture diffuses through the membrane into the electrolyte film, where it equilibrates with HCO3- ions and shifts the buffer's pH according to the carbonic-acid equilibrium. An internal pH glass electrode reads that pH shift, and the sensor's electronics convert it to a partial-pressure CO2 reading using a logarithmic relationship.

The dominant commercial Severinghaus probe in bioprocess is the Mettler Toledo InPro 5000i, which is integrated with Mettler's Intelligent Sensor Management (ISM) digital platform for predictive diagnostics. The InPro 5000i covers a 0–1000 mbar pCO2 range, has a t90 response time under 120 seconds at 25°C, and is autoclavable and SIP-compatible up to 130°C. The membrane body is field-replaceable on the existing probe shaft, which is the design feature that has kept it competitive against newer optical alternatives — operators can refurbish a probe in place rather than send it out for service.

When Severinghaus wins

Severinghaus dominates three scenarios. First, existing cGMP installed base — a typical commercial mammalian facility built between 2005 and 2020 has Severinghaus probes plumbed into every bioreactor port, validated against the InPro 5000i datasheet. Switching all of them to optical would trigger an IQ/OQ/PQ revalidation campaign that costs more than 5–10 years of electrolyte refills. Second, microbial fermentation at high pCO2 (above 200 mbar) — the InPro 5000i's 0–1000 mbar range covers high-density E. coli and yeast fed-batch where dCO2 routinely climbs above 300 mbar; many fluorescence optical sensors saturate or lose accuracy above 200 mbar. Third, academic and pilot-scale teaching — the lower capital cost per channel and the ubiquity of Mettler's training material make Severinghaus the default in research-grade glass bioreactors.

How optical CO2 sensors work

"Optical CO2 sensor" is a category, not a single technology. Two physical principles share the label, and they have very different operating envelopes.

Mid-IR absorption (solid-state). CO2 has a strong absorption band near 4.26 µm. A small mid-IR LED inside the probe emits at this wavelength, the light passes through a polymer-protected sample volume in contact with the culture, and a photodiode measures the transmitted intensity. The Beer-Lambert relationship between transmittance and CO2 partial pressure gives the reading. This is the principle Hamilton CO2NTROL uses. Because there is no electrolyte and no consumed reagent, the probe is solid-state and maintenance-free — Hamilton specifies it as having no consumables for the rated lifetime. The CO2NTROL covers 5–1000 mbar, mounts in standard PG13.5 bioreactor ports, and is SIP- and CIP-compatible up to ~140°C.

Fluorescence quenching of an HPTS dye in a bicarbonate buffer. This is the original optical Severinghaus principle. A pH-sensitive fluorescent dye (typically 1-hydroxypyrene-3,6,8-trisulfonate, HPTS) is dissolved in a thin bicarbonate buffer film behind a CO2-permeable membrane, with an LED exciting the dye and a photodiode measuring its fluorescence. CO2 from the culture diffuses through the membrane, equilibrates with the buffer, shifts its pH, and changes the dye's fluorescence. PreSens uses Dual Lifetime Referencing (DLR) — a second inert reference dye in the same matrix — to internally correct for intensity drift and ageing of the LED. PreSens offers HPTS-based optical CO2 in three formats: insertable benchtop probes (DP-CD1), gamma-sterilised SP-CD1 sensor spots for single-use bags, and FTC-SU-CD1 single-use flow-through cells for perfusion bypass loops.

When optical wins

Mid-IR optical wins on lifetime maintenance for new stainless-steel cGMP builds — no electrolyte to refill, no membrane to replace, no consumables to qualify, no operator hours per probe per year. Hamilton estimates a roughly 3-year crossover where the avoided maintenance overtakes the higher capital cost. HPTS-dye optical wins decisively on single-use bag compatibility — the PreSens spot is glued to the inside of the bag wall by the bag vendor, gamma-sterilised together with the bag, and read non-invasively through a transparent window. A traditional Severinghaus form factor with a refillable electrolyte chamber simply cannot integrate with a sealed bag. Optical is also preferred in perfusion and small-scale parallel bioreactors (pCO2 mini FTC-CD1, Sartorius ambr) where the small-volume probe footprint matters.

Pros and cons

Severinghaus electrode

Advantages

  • Industry standard since 1957 — deep validation heritage
  • Wide working range (0–1000 mbar) covers microbial high-pCO2
  • Lower capital cost per channel than mid-IR optical
  • Field-replaceable membrane body (InPro 5000i) — refurbish in place
  • Mature ISM digital diagnostics in Mettler ecosystem

Disadvantages

  • Bicarbonate electrolyte requires periodic refilling
  • Membrane fouling and pinhole leaks are routine failure modes
  • Calibration drift between membrane changes
  • No single-use bag form factor exists — incompatible with SUBs
  • 10–20 operator-hours per probe per year for membrane/electrolyte service

Optical CO2 sensor

Advantages

  • No electrolyte, no membrane changes — maintenance-free (mid-IR)
  • Single-use bag-compatible (HPTS spots glued in by bag vendor)
  • Higher precision in physiological pCO2 range (40–80 mmHg)
  • Lower long-term calibration drift (solid-state, no electrolyte dilution)
  • Pre-calibrated from the factory — minimal pre-batch work

Disadvantages

  • 1.3–1.5× higher capital cost per channel (mid-IR)
  • Smaller installed base — younger validation track record
  • HPTS dye saturates above ~250 hPa — limited for microbial high-pCO2
  • Single-use spots are consumables tied to the bag (per-batch cost)
  • Retrofit into validated process triggers IQ/OQ/PQ revalidation

Which should you choose?

The CO2 sensor decision is dominated by vessel type and process modality, with capital budget as a secondary consideration.

New stainless-steel cGMP facility

Reusable bioreactors at 200–10,000 L for commercial mAb or recombinant protein. Solid-state mid-IR (Hamilton CO2NTROL) eliminates electrolyte refills across 100+ batches per year and integrates with the Hamilton Arc digital ecosystem alongside VisiFerm DO and EasyFerm pH.

Choose Optical (mid-IR)

Single-use bag bioreactor

Sartorius Biostat STR, Thermo HyPerforma, Cytiva Xcellerex, or Pall Allegro at 50–2,000 L. PreSens SP-CD1 spots are glued inside the bag at manufacture and gamma-sterilised together with the bag. Severinghaus electrodes have no equivalent form factor.

Choose Optical (HPTS)

High-density microbial fermentation

High-OUR E. coli or Pichia with dCO2 climbing above 200–300 mbar during induction. Severinghaus (InPro 5000i) covers the 0–1000 mbar range cleanly; HPTS-dye optical can saturate at the upper end.

Choose Severinghaus

Validated legacy cGMP process

Existing commercial process running InPro 5000i for the last 10+ years. Switching probe technology triggers IQ/OQ/PQ revalidation, SOP updates, and cleaning validation refresh — usually more expensive than continuing with electrolyte refills.

Keep Severinghaus

Real-world use cases

Four representative deployments and why each team converged on their choice.

CHO mAb fed-batch · 2,000 L SS
Mid-IR optical (Hamilton CO2NTROL)

New build with full Hamilton Arc ecosystem (VisiFerm DO, EasyFerm pH, CO2NTROL). One-time pre-run air-zero calibration, RS485 to DeltaV, no consumable PO line for the lifetime of the probe. Maintenance overhead matched to DO and pH probes — single ISM-style diagnostic dashboard.

E. coli high-density · 100 L SS
Severinghaus (Mettler InPro 5000i)

dCO2 climbs to 350–500 mbar during oxygen-limited induction with continuous glucose feed. The InPro 5000i's 0–1000 mbar range covers the regime cleanly. Membrane refurbished every 6 months, electrolyte refilled at each campaign turnaround.

AAV HEK293 · 200 L SU bag
Optical spot (PreSens SP-CD1, bag-integrated)

Spot pre-glued inside the Sartorius Biostat STR 200 bag, gamma-sterilised together with the bag, ready at installation. Read by an external pCO2 mini optical reader through the transparent window. No Severinghaus form factor was ever in consideration.

CHO perfusion · 50 L bypass loop
Optical flow cell (PreSens FTC-SU-CD1)

Single-use beta-irradiated flow-through cell in a perfusion bypass with C-Flex tubing. Reads pCO2 in the recirculation stream without breaching the main vessel. Perfusion dCO2 routinely accumulates above 100 mmHg, and the FTC's 8–180 mmHg range covers it.

Not sure which CO2 sensor fits your scale and modality?

Answer a few quick questions about scale, vessel, parameters, and budget, and get ranked sensor recommendations covering CO2, DO, pH, biomass, and PAT analytics.

Open the Sensor Selection Tool

Cost and lifecycle considerations

Total cost of ownership inverts the capital-cost story

Severinghaus wins on probe-price alone, but consumables (electrolyte cartridges, replacement membrane bodies, calibration gases) plus operator time for membrane swaps converge the 3-year TCO with mid-IR optical. The decisive factor is operator hours — each membrane change on an InPro 5000i takes 2–4 hours including re-calibration. At 100 batches/year with a change every 4–6 months, that is 4–12 operator-hours per probe per year that solid-state optical eliminates entirely. For single-use bag deployments the comparison is moot — the per-batch consumable cost of an SP-CD1 spot is bundled into the bag price, not a separate line item.

A single 2,000 L mammalian bioreactor deploying solid-state mid-IR CO2 (Hamilton CO2NTROL) costs approximately £4,500–£7,500 for the probe with integrated transmitter, with effectively zero recurring consumable cost over the rated lifetime. The same vessel with a Severinghaus InPro 5000i costs £3,000–£5,000 up front but adds £400–£900/year in consumables (electrolyte cartridges, membrane bodies, calibration gas refills) plus 4–12 hours/year of operator time for replacements.

For single-use deployments the calculation is different. The PreSens SP-CD1 spot itself is inexpensive (~£15–£40 per spot), but it ships pre-integrated in the bag at a per-batch cost rolled into the bag price. The pCO2 mini reader unit (£3,000–£5,000) is reused across batches. Over a 100-batch single-use campaign, the all-in single-use CO2 monitoring cost is comparable to a 2-channel reusable Severinghaus deployment, but with the upside that there is no probe to clean, recalibrate, or revalidate between batches.

Cost componentSeveringhaus (InPro 5000i)Optical mid-IR (CO2NTROL)
Probe + transmitter (per channel)£3,000–£5,000£4,500–£7,500
Consumables / year (electrolyte, membrane, cal gas)£400–£900~£0–£100 (cal gas only)
Operator labour / year (maintenance)4–12 hours<1 hour
3-year TCO (per channel, 100 batches/yr)£4,500–£7,500£4,800–£7,800

Vendor landscape

The dissolved CO2 sensor market is more concentrated than the DO market — three vendors cover essentially the entire bioprocess installed base.

Severinghaus electrode vendors

Optical CO2 sensor vendors

Frequently asked questions

What is the difference between a Severinghaus and an optical CO2 sensor?
A Severinghaus sensor is a potentiometric electrode: CO2 diffuses through a gas-permeable membrane into a thin bicarbonate electrolyte film, shifts its pH, and an internal pH glass electrode reads that shift. Optical CO2 sensors come in two flavors. The first uses the same Severinghaus chemistry but reads the pH of the buffer with a fluorescent dye (HPTS) instead of a glass electrode — this is the principle behind PreSens spots. The second is solid-state mid-IR absorption, which detects CO2 directly without any buffer or consumables — this is what Hamilton CO2NTROL uses. The practical consequence is that potentiometric Severinghaus sensors have an internal electrolyte that has to be refilled and a membrane that ages, while solid-state optical sensors are maintenance-free for the life of the probe.
Is the Mettler InPro 5000i a Severinghaus sensor?
Yes. The Mettler Toledo InPro 5000i is a potentiometric Severinghaus CO2 sensor. CO2 diffuses through a silicone-reinforced PTFE membrane into a bicarbonate inner electrolyte, where it equilibrates with HCO3- ions and changes the buffer pH. An internal pH electrode measures that change and converts it to a pCO2 reading. The sensor has a measuring range of 0–1000 mbar pCO2, an accuracy of ±10% across most of that range, and a t90 response time under 120 seconds at 25°C. It is autoclavable and CIP/SIP compatible up to 130°C.
Why do optical CO2 sensors work better for single-use bioreactors?
Single-use bioreactors are sealed, gamma-irradiated bags that cannot accommodate a probe with an internal liquid electrolyte and a refillable membrane assembly — both would breach sterility. Optical CO2 sensors with the dye on a small adhesive spot or in a flow-through cell can be pre-integrated by the bag manufacturer at assembly, gamma-sterilised together with the bag, and read non-invasively through a transparent window. PreSens SP-CD1 spots and the FTC-SU-CD1 single-use flow-through cell are designed exactly for this workflow. A traditional Severinghaus probe simply has no equivalent single-use form factor.
How accurate are optical CO2 sensors compared to Severinghaus electrodes?
Both can hit ±5–10% in their working range under good calibration. Optical fluorescence sensors using HPTS dye (PreSens SP-CD1) report ±0.5 mmHg at 15 mmHg pCO2 and ±1.2 mmHg at 45 mmHg pCO2 — high precision in the physiological mammalian range (40–80 mmHg). Severinghaus probes (InPro 5000i) cover a wider range (0–1000 mbar) at ±10%, which makes them more useful for microbial fermentation where dCO2 routinely exceeds 200 mbar. Solid-state mid-IR optical sensors (Hamilton CO2NTROL) advertise comparable accuracy across 5–1000 mbar with no consumable drift. In practice the dominant error in any CO2 reading is calibration drift, not the sensor's nominal accuracy spec.
What is the response time of CO2 sensors in bioreactors?
All in-line CO2 sensors are slow compared with DO probes because CO2 has to diffuse through a gas-permeable membrane and equilibrate with an internal phase. Severinghaus electrodes (Mettler InPro 5000i) have a t90 of 60–120 seconds at 25°C. Optical Severinghaus sensors with HPTS dye (PreSens) sit in the same window. Solid-state mid-IR optical sensors (Hamilton CO2NTROL) are similar at 60–180 s because the rate-limiting step is still diffusion through the polymer membrane that protects the optical path. For closed-loop dCO2 control this is fine — CO2 dynamics in a bioreactor evolve over minutes, not seconds.
Can optical CO2 sensors be sterilized in place (SIP)?
Yes for solid-state mid-IR sensors. Hamilton CO2NTROL is SIP- and CIP-compatible at up to ~140°C, with EHEDG-approved hygienic design and stainless steel wetted parts polished to Ra <0.4 µm. Optical fluorescence sensors with HPTS dye are mostly used in autoclavable benchtop probes (PreSens DP-CD1) or single-use spot/flow-through formats that are gamma-sterilised by the bag vendor — neither is in-place steam sterilised in the same way as a stainless-steel reusable bioreactor probe. For commercial-scale stainless-steel mAb fed-batch with SIP cycling, a hygienic Severinghaus or solid-state mid-IR optical sensor is the realistic choice.
Do CO2 sensors need calibration?
Yes. Severinghaus electrodes are calibrated by exposure to two known CO2 partial pressures — typically room air (~0.4 mbar pCO2) and a 5–10% CO2 reference gas — before each batch and after every electrolyte refill. Calibration drift between cycles is a known weakness, which is why Mettler ISM diagnostics track membrane health and electrolyte life. Optical fluorescence sensors come pre-calibrated from the factory at 20°C, 37°C, and physiological osmolality, and only need a 1-point adjustment before use. Solid-state mid-IR sensors are factory-calibrated for the life of the probe, with a recommended air-zero check before installation.
Which CO2 sensor is best for CHO mAb fed-batch?
For new commercial CHO mAb processes in stainless steel bioreactors, solid-state mid-IR optical CO2 sensors (Hamilton CO2NTROL) and Severinghaus electrodes (Mettler InPro 5000i) are the two viable choices. CO2NTROL wins on lifetime maintenance — no electrolyte refills, no membrane changes, ISM-style diagnostics — at a higher capital cost. InPro 5000i wins on installed base and validation heritage, with most existing Sartorius, ABEC, and legacy Pierre Guérin lines already plumbed for it. For CHO in single-use bags (Sartorius Biostat STR, Thermo HyPerforma, Cytiva Xcellerex), PreSens optical spots or flow-through cells are essentially the only option because no Severinghaus form factor integrates with a sealed gamma-sterilised bag.

Resources and references