Chemostat Steady State — µ = D = F/V

This is the strangest true statement in bioprocess engineering, and it is worth stating flatly before explaining it: in a steady-state chemostat, the operator sets the specific growth rate of a living culture by turning a pump. The organism does not get a vote.

Here is why. A chemostat is a stirred vessel fed sterile medium at a constant flow $F$, with an overflow that holds the working volume $V$ constant. Everything fed in must leave, cells included, and cells leave at the fractional rate $D = F/V$ per hour — the dilution rate, which is simply the reciprocal of the mean residence time. For the biomass in the vessel to hold steady, growth must exactly replace that loss. The biomass balance is $dX/dt = \mu X - DX$, and setting it to zero gives $\mu = D$. Nothing about the organism enters that argument anywhere. Two entirely different species in two identical chemostats at the same pump setting grow at exactly the same specific rate.

What differs between them is the residual substrate each needs to sustain it. The culture cannot change $\mu$ — the pump has fixed it — so it adjusts the only variable left to it, and consumes substrate down to whatever concentration makes Monod's equation balance at that $\mu$. Run the pump faster and the residual substrate RISES, because a faster-growing culture needs a richer environment. Run it slower and the residual falls, sometimes to micrograms per litre. That is why a chemostat is the classic enrichment device: held at a low dilution rate on one substrate, it selects ruthlessly for whichever organism can grow at that rate on the least substrate — the lowest $K_s$ wins, regardless of who is faster at the top end.

Monod and, independently, Aaron Novick and Leo Szilard both published the chemostat in 1950. Its importance is methodological. In a batch culture, growth rate and substrate concentration and biomass and product all change together, so nothing can be held constant while something else is varied — every measurement is a moment on a moving curve. In a chemostat, growth rate is a dial on the front of a pump and it stays where you put it for weeks. That is what makes it possible to ask, and answer, what a cell does differently at 0.1 h⁻¹ than at 0.4 h⁻¹, which turns out to be a great deal: cell size, ribosome content, storage compounds, product spectrum and the expression of hundreds of genes all shift with growth rate.

The chemical-engineering reader will recognise the vessel as a CSTR, and $1/D$ as its space time $\tau$. The recognition is exact, and the only difference is that the catalyst reproduces itself. The wastewater reader will recognise the same balance under a different name: the mean cell residence time, or sludge age. There the two are deliberately decoupled — solids are settled out and returned, so the cells stay far longer than the water does — which is precisely how a treatment plant retains slow-growing nitrifiers at a hydraulic residence time that would wash them out in an hour.

Three practical cautions. Steady state takes time to arrive: allow four or five residence times after any change before you sample, or you will be measuring a transient and calling it an equilibrium. Wall growth ruins the argument, because a biofilm on the glass is not being washed out at $D$ and quietly supplies the vessel with cells it did not have to grow; a long chemostat run needs the walls kept clean. And the feed must be genuinely sterile, since a contaminant that grows faster than $D$ on the same substrate will take the vessel over, and the chemostat will select for it with the same efficiency it selects for anything else.