Sensors
Portable biomass analyzers
Why optical density goes nonlinear once scattering dominates, how capacitance discriminates living cells by intact membranes, why the dry-weight anchor drifts with morphology, and what calibration models do for portable NIR composition analysis.
Biomass is a mass, but no field instrument measures a mass of living cells directly. Every number called biomass — in a fermenter, a bioreactor sample, a harvested slurry — is a proxy plus a transfer function mapping that proxy to grams per litre. The instruments differ in which physical property they borrow, and the honest comparison between them is where each transfer function breaks.
Optical density and its ceilings
Attenuation of light by a culture has two components: true absorption and scattering. The Beer–Lambert law is linear only for the first case, with each particle independently removing photons from the beam. Scattering does not remove photons — it redirects them, and once the suspension is dense enough for a photon to be scattered repeatedly, path lengths stretch and attenuation grows sublinearly with concentration. Past that regime, dilution is not a convenience but a measurement requirement. Even below the ceiling, optical density is a relative quantity: it depends on wavelength, on particle size distribution, on cell shape and on the refractive-index contrast between cell and medium. An OD unit is specific to the instrument and the organism, and a correlation between OD and dry weight has to be built per strain and per growth phase — it is not transferable luggage.
Capacitance: counting intact membranes
The complementary proxy borrows a dielectric property instead. Electrodes across the suspension apply a radio-frequency field; intact plasma membranes are thin dielectrics impermeable to ions, so each living cell polarizes like a tiny capacitor, while lysed cells, membrane ghosts, debris and gas bubbles contribute only conductivity. Around the megahertz range this separation — the beta-dispersion — makes the measured capacitance a measure of intact-membrane volume: live-cell discrimination by physics, not by dye exclusion. The transfer function has its own conditions. Background conductivity of the medium must be compensated, and the capacitance-to-mass ratio depends on membrane area per unit biovolume — a filamentous organism or a heavily vacuolated cell shifts it without any change in viable mass.
The dry-weight anchor and its morphology problem
Dry weight is the anchor everyone calibrates against, and it is honest — but the correlation that other methods build on it breaks with morphology. Filaments and single cells of the same mass scatter light differently and present different membrane area. Vacuoles and storage compounds change mass without changing cell count, so a culture drifting into stationary phase walks away from any exponential-phase calibration even though the instrument is working perfectly. This is not a defect to fix but a property to respect: the mapping between proxy and mass tracks physiological state, which is exactly the thing that changes fastest during a run.
Portable NIR analyzers are a different proxy again: they report composition — moisture, protein, oil — of harvested or feedstock biomass through calibration models, not cell mass. They are model-bound like every other method on this page, and the model is the instrument. What all these share is the chain of continuous monitoring — recognition layer, transducer, signal — with one addition: in biomass measurement the last stage is a statistical transfer function, and its validity domain, not the optics, is usually the binding constraint.