A biomass sensor turns the most basic question in a bioreactor run — how many cells do I have right now? — into a continuous, real-time signal. Instead of pulling a sample every few hours and counting cells offline, an in-line biomass probe or a software model reports cell concentration second by second, so you can see growth rate live, trigger a feed or induction at a target density, and catch the death phase the moment it starts. This guide explains the main biomass sensor types, what each one actually measures, their detection limits, and how to choose the right one for your process.
What is a biomass sensor?
A biomass sensor is an instrument that measures cell concentration in situ, in the running culture, without removing a sample. The critical distinction between every biomass monitor on the market is which biomass it reports: viable cells only, or all cells.
Viable biomass is the concentration of living, membrane-intact cells — the cells actually making your product. Total biomass is every cell plus debris, alive or dead. In early exponential growth the two track together, but once the death phase begins they diverge sharply, and that gap is itself a powerful viability readout.
Biomass sensors fall into three families: capacitance/dielectric probes (viable), optical probes (total), and soft sensors that infer biomass from other measurements. Offline methods such as OD600 and dry cell weight are not sensors in the in-line sense, but they remain the reference every probe is calibrated against.
Capacitance / dielectric biomass sensors
A capacitance biomass sensor measures viable cell density using dielectric spectroscopy: in a radio-frequency alternating field, viable cells with intact plasma membranes polarise like tiny capacitors, and the measured permittivity is proportional to the viable biomass volume in the field. Dead cells and debris have leaky membranes, cannot sustain charge separation, and are effectively invisible — which is exactly why the technique is prized.
Dielectric spectroscopy is now the most common method for estimating live cell concentration in mammalian cell culture. Aber Instruments pioneered the commercial instrumentation; Hamilton's Incyte is the other widely deployed line. A subtlety worth knowing: because cells are not uniform in size, the raw measurement is really a viable volume, not a viable count — conversion to cells/mL depends on cell-specific parameters such as mean cell radius, internal conductivity and membrane capacitance, which drift through a culture and are the main source of model error.
Worked example: permittivity to viable cell density
A capacitance probe reads a permittivity change of Δε = 8.0 pF/cm at the measurement frequency. During calibration, paired offline counts gave a linear fit:
VCD (106 cells/mL) = 2.4 × Δε (pF/cm) − 0.6
Substituting the live reading:
VCD = 2.4 × 8.0 − 0.6 = 18.6 × 106 cells/mL
The slope (2.4) is the cell-line-specific cell constant; it must be re-fitted when you change cell line, scale, or medium, because mean cell size sets how much permittivity each viable cell contributes.
Strengths: viable-only signal, robust to gas bubbles and colour, and a real-time growth rate you can act on. Limits: a detection floor around 0.5–1 × 106 cells/mL for mammalian cultures (below which signal-to-noise degrades), and the cell-size dependence that makes the calibration non-transferable. For a head-to-head against optical, see our capacitance vs optical biomass sensor comparison, and the vendor-specific Hamilton Incyte review and Aber FUTURA review.
Not sure which probe fits your bioreactor?
Match measurement type, scale, and single-use needs to a sensor in a couple of clicks.
Optical biomass sensors (turbidity / NIR)
An optical biomass sensor measures total cell density from the turbidity of the culture: near-infrared (NIR) light is sent into the broth, and every particle that scatters or absorbs it — viable cells, dead cells, debris, even some precipitates — contributes to the signal. Because it counts everything, optical density correlates strongly with total biomass but poorly with viability.
Two optical geometries dominate. Transmission probes measure light lost across a fixed path length and suit low-to-moderate densities. Backscatter probes measure light reflected back from particles and stay linear to much higher densities — important for high-cell-density microbial fermentation, where a transmission signal saturates. NIR (around 850–880 nm) is chosen because medium components absorb less there than in the visible, so the signal tracks cells rather than broth colour.
Optical sensors shine where capacitance struggles: very high density, microbial speed, and total-biomass harvest decisions. Their weakness is the inability to distinguish live from dead, plus sensitivity to bubbles and to anything else that scatters light. Running an optical probe alongside a capacitance probe is a common, powerful combination: the viable signal (capacitance) and total signal (optical) are read together, and viability is inferred from the widening gap between them.
Figure 2. In a fed-batch, viable (capacitance) and total (optical) biomass overlap during growth and diverge in the death phase — the gap is a real-time viability readout.
Soft sensors and model-based biomass
A soft sensor estimates biomass from measurements you already have — off-gas (OUR/CER), pH and base addition, dissolved oxygen, feed rate, and agitation — through a mass-balance or data-driven model, with no extra probe in the vessel. The classic example is biomass from the oxygen uptake rate: in aerobic culture, cumulative oxygen consumed is tightly coupled to cells grown, so an off-gas analyser plus a yield coefficient gives a continuous biomass estimate for the price of software.
Soft sensors are attractive because they add no consumable and no port, work in any vessel including ones too small for a probe, and can fuse several signals for robustness. The trade-off is that they are only as good as the model and its calibration: a wrong yield coefficient or an un-modelled metabolic shift propagates straight into the biomass estimate. They pair naturally with hard sensors — a capacitance probe to anchor the model, a soft sensor to extend it. Our deep dive on soft sensors in bioprocess covers model building and validation, and PAT bioprocess monitoring sets the regulatory context.
Offline reference methods
Every in-line biomass sensor is only as trustworthy as the offline reference it is calibrated against. The three references in routine use each measure a different thing, which is why a probe calibrated to one will not read another correctly:
- OD600 — optical density at 600 nm, fast and cheap, a total-biomass proxy for microbial work. Convert OD to cells or dry weight with an OD600 calculator.
- Dry cell weight (DCW) — gravimetric, the gold standard for total biomass in g/L, but slow (filter, wash, dry, weigh).
- Viable cell density (VCD) — trypan-blue or automated viable counts, the reference for capacitance probes.
Pick the reference that matches what the sensor physically measures: DCW or OD600 for optical, VCD for capacitance. Mismatching them is the single most common cause of a "drifting" biomass calibration. To turn calibrated readings into growth rate and doubling time, feed them into a growth curve fitter.
Which biomass sensor measures what
The table below summarises the practical differences. The right column — what each sensor actually measures — is the one that should drive your choice.
| Type | Measures | Principle | Useful range | Single-use | Best for |
|---|---|---|---|---|---|
| Capacitance / dielectric | Viable biomass | RF permittivity of intact membranes | ~0.5–1 ×106 to >108 cells/mL | Yes (probe or patch) | VCD, physiology, induction/feed triggers |
| Optical — transmission | Total biomass | NIR light loss over path | Low–moderate density | Yes | Mammalian total density, clarification cues |
| Optical — backscatter | Total biomass | NIR light reflected from particles | Up to very high density | Yes | High-density microbial fermentation |
| Soft sensor | Viable or total (model-dependent) | Mass balance / ML from OUR, pH, feed | Whole run | N/A (software) | No-probe vessels, signal fusion |
| Offline (OD600 / DCW / VCD) | Reference | Photometry / gravimetry / counting | Method-dependent | N/A | Calibration ground truth |
Figure 3. Each sensor correlates with the biomass it physically measures — and not with the other. Choose the sensor whose physics matches your question.
How to choose a biomass sensor
Start from the decision you need the signal to support, not from the sensor spec sheet. The flow below resolves most cases in three questions: what biomass do you care about, how dense is the culture, and is there a probe port and budget?
Three more practical filters once the type is chosen: single-use compatibility (a pre-irradiated probe or patch sensor if you run single-use bags), scale and port availability (microbioreactors may only support a soft sensor), and calibration burden (capacitance needs a fresh cell constant per cell line; optical is more transferable). When you have shortlisted a type, the sensor selection tool maps it to specific products, and the capacitance vs optical comparison resolves the most common fork.
Turn a calibrated biomass signal into growth rate
Fit your in-line cell-density data to get µ, doubling time, and phase boundaries.
Frequently Asked Questions
What is a biomass sensor?
A biomass sensor is an in-line probe or soft-sensor model that measures cell concentration in a bioreactor in real time, without taking a sample. Capacitance (dielectric) sensors report viable cell density from the permittivity of intact cell membranes, while optical sensors report total cell density from near-infrared light scattering.
How does a capacitance biomass probe work?
In a radio-frequency electric field, viable cells with intact membranes polarise like tiny capacitors. The probe measures the resulting permittivity (capacitance per unit area), which is proportional to viable biomass volume. Dead cells and debris have leaky membranes, cannot hold charge, and are therefore invisible to the sensor — so capacitance tracks viable, not total, biomass.
What is the difference between viable and total biomass measurement?
Viable biomass counts only living, membrane-intact cells and is best measured by capacitance/dielectric spectroscopy (correlates with viable cell density, R² ≈ 0.75). Total biomass counts all particles, living or dead, and is measured by optical turbidity (correlates with total cell density, R² ≈ 0.83). Running both lets you track viability as the gap between the two signals widens in the death phase.
Can biomass sensors be used in single-use bioreactors?
Yes. Both capacitance and optical biomass sensors come in single-use formats — pre-gamma-irradiated probes through a sealed port, or patch/window sensors read by a reusable external transmitter. Single-use removes cleaning validation and cross-contamination risk but adds a small per-batch consumable cost.
What does a biomass monitor measure during a fermentation?
A biomass monitor outputs a continuous signal proportional to cell concentration — permittivity in pF/cm for capacitance, or a turbidity/absorbance unit for optical. Calibrated against an offline reference (DCW, OD600, or VCD), it gives real-time growth rate, lets you trigger feeds or induction at a set density, and flags the onset of the death phase.