When the medium defeats the proxy
Getting accurate growth data is important in every cultivation setting, whether the work is process development, applied research, biorefinery, anaerobic biology, or industrial fermentation. The fastest and most ubiquitous readout is optical density at 600 nm: cheap, fast, available on every benchtop. In clean broths it works. In some media, it does not. The matrix itself defeats the measurement, and no amount of careful blanking will rescue it.
Anaerobic broths with resazurin shift color during cultivation, putting the indicator's optical state directly in the OD600 measurement window. Lignocellulosic hydrolysates carry undigested cellulose and lignin particulates that scatter at 600 nm independently of the cells. Low-cost agro-industrial feedstocks like molasses and corn steep liquor, the staple carbon and nitrogen sources of bulk bioproduction, combine deep visible-range color from melanoidins with particulates and lot-to-lot variability that no blanking strategy can subtract. In each, the optical-density signal carries everything in the cuvette except a clean count of the cells. If your medium is one of these, or behaves like one, this article is about an alternative.
Most of the published OD600 literature is written about clear minimal media: M9 plus glucose, defined glycerol broths, standard LB. That is not where most applied research happens. The matrices that matter for process development, fermentation scale-up, biorefinery work, and anaerobic research are richer, dirtier, and harder to read optically. When OD600 doesn't work, scientists usually fall back on slow methods such as CFU plating or dry cell weight, or accept a noisy signal and post-rationalize the curve. Neither delivers growth data trustworthy enough for decisions that depend on knowing what the culture is doing right now.
The rest of this article covers three media families where OD600 fails for well-characterized reasons, why direct impedance-based cell counts sidestep the problem, and where the limits of the BactoBox® alternative actually lie.
Common media where OD600 breaks down
| Medium / matrix | Common applications | What defeats OD600 |
|---|---|---|
| Anaerobic media with resazurin (RCM, PYG, Wilkins–Chalgren, mGAM, Schaedler) | Clostridium, Bacteroides, Faecalibacterium, gut-microbiome research, dental microbiology, rumen microbiology | Resazurin absorbs at 600 nm; trace oxygen during sampling back-shifts the color, placing the indicator's optical state directly in the measurement window[1]. |
| Lignocellulosic hydrolysates | Biorefinery, second-generation ethanol, lignocellulosic biomass conversion. Organisms include Saccharomyces, Zymomonas mobilis, Clostridium, and mixed cultures | Cellulose fines and lignin particulates scatter at 600 nm independently of the cells; OD600 is not applicable for cultures with insoluble particles[2]. |
| Low-cost agro-industrial feedstocks (molasses, corn steep liquor, crude glycerol) | Bulk bioproduct manufacturing: baker's yeast, bioethanol, amino acid fermentation (L-glutamate, L-lysine via Corynebacterium glutamicum), citric acid (Aspergillus niger), lactic acid, recombinant E. coli and Bacillus processes | Molasses carries melanoidins (Maillard products) and other colored impurities that absorb broadly across the visible range; raw molasses routinely requires decolorization pretreatment before fermentation[3]. Corn steep liquor and similar byproducts add substantial lot-to-lot variability in turbidity and composition[4]. |
The common shape across these three is that the failure is not about the cells. It is about everything else in the cuvette: the indicator, the particulates, the precipitates, the pigments, the additives. Calibration cannot rescue this, because every one of those non-cell contributors depends on the medium, the lot, the cultivation conditions, the redox state of the broth, and the timing of the sample.
Some growth-data failures look matrix-driven but are actually organism-driven. Mycoplasma species, for example, produce so little turbidity at peak growth that OD600 fails regardless of which broth they are cultivated in. The issue is the cell size and wall-less envelope, not the medium. The dedicated Mycoplasma growth data article covers that case.
Why OD600 fails in these media
OD600 is an extinction measurement, not a cell count. The signal at 600 nm sums every contributor in the optical path that removes light from the beam, whether by scattering (cells, peptone fines, precipitates, antifoam droplets) or by absorbance (indicator dyes, melanoidins from caramelized sugars, other colored impurities)[5]. In a clean broth, cells dominate and OD600 is a reasonable cell-number proxy. In the media in the table above, the non-cell contributors are either too large to ignore, too variable to subtract, or too entangled with the cells over time to separate.
Calibration cannot rescue this. A calibration that works for one batch of hydrolysate or one lot of corn steep liquor will fail on the next, because the non-cell contribution to OD changes lot-to-lot, run-to-run, and even sample-to-sample as additions, oxygen exposure, and pH corrections shift the optical state of the broth.
What BactoBox® counts in these media
BactoBox® uses impedance flow cytometry. Each particle that crosses a pair of microelectrodes in a microfluidic channel is recorded as a single event, with an electrical signature determined by its envelope and cytoplasm[6]. A cell with an intact lipid membrane and a conductive cytoplasm is classified as a bacterial cell. A non-cell particle, such as a peptone fragment, a precipitate crystal, or a resazurin molecule, has a different impedance signature and is classified separately. The measurement does not depend on the optical properties of the broth. The same cells get counted whether the medium is clear, opaque, colored, or shifting through the cultivation.
Understanding BactoBox® cell counts covers the principle, the classification logic, and the comparison with other methods in detail. BactoBox® cell counts have been benchmarked against colony-forming-unit plating across six bacterial genera in fermentation processes, with near-perfect correlation through exponential, deceleration, and stationary phases[7].
The 0.5–5 µm calibration range still applies. In matrices with substantial particulate content sitting inside that gate, a brief check against a reference method on a new medium remains good practice. We are happy to help work through that for a specific application.
What direct counts unlock
In matrices where OD600 is unusable, scientists typically choose between two bad workflows. Either accept slow downstream methods like CFU plating or dry cell weight at the end of the run, and design experiments around the readout latency. Or invest in elaborate sample-prep workarounds: multi-step dilutions, filter blanks, regression curves against CFU, lot-by-lot recalibration, switching to NIR or capacitance probes that work in the bioreactor but not at the lab bench.
A direct cell count in roughly two minutes per sample sidesteps both. Process development can happen in the medium the process actually runs in, with the same readout time as a clean broth. Strain comparisons in anaerobic or hydrolysate media become direct cell-density comparisons rather than turbidity-difference comparisons. Yield and productivity calculations are built on cells per milliliter, which is the unit downstream economics depend on already. Phase boundaries (lag, exponential, deceleration, plateau) become observable in real time, in the matrix where the process lives, on the same instrument across applications. Growth data you can trust comes from counting cells directly, not from interpreting how much light got through the broth.
References
- O'Brien, J., Wilson, I., Orton, T., & Pognan, F. (2000). Investigation of the Alamar Blue (resazurin) fluorescent dye for the assessment of mammalian cell cytotoxicity. European Journal of Biochemistry, 267(17), 5421–5426. https://doi.org/10.1046/j.1432-1327.2000.01606.x
- Duedu, K. O., & French, C. E. (2017). Data for discriminating dead/live bacteria in homogenous cell suspensions and the effect of insoluble substrates on turbidimetric measurements. Data in Brief, 12, 169–174. https://doi.org/10.1016/j.dib.2017.04.003
- Roukas, T. (1998). Pretreatment of beet molasses to increase pullulan production. Process Biochemistry, 33(8), 805–810. https://doi.org/10.1016/S0032-9592(98)00048-X
- Wahjudi, S. M. W., Petrzik, T., Oudenne, F., Lera Calvo, L., & Büchs, J. (2023). Unraveling the potential and constraints associated with corn steep liquor as a nutrient source for industrial fermentations. Biotechnology Progress, 39(6), e3386. https://doi.org/10.1002/btpr.3386
- Myers, J. A., Curtis, B. S., & Curtis, W. R. (2013). Improving accuracy of cell and chromophore concentration measurements using optical density. BMC Biophysics, 6, 4. https://doi.org/10.1186/2046-1682-6-4
- Bertelsen, C. V., Skands, G. E., González Díaz, M., Dimaki, M., & Svendsen, W. E. (2023). Using impedance flow cytometry for rapid viability classification of heat-treated bacteria. ACS Omega, 8(8), 7714–7721. https://doi.org/10.1021/acsomega.2c07357
- Jordal, P. L., González Díaz, M., Aalund, F., & Skands, G. (2025). Performance qualification of impedance flow cytometry as a rapid in-process control proxy for colony-forming units in bacterial fermentation processes. Journal of Microbiological Methods, 238, 107284. https://doi.org/10.1016/j.mimet.2025.107284