Chemostat Washout (Critical Dilution Rate)

There is one limit on the chemostat's remarkable property, and this is it. Setting $\mu = D$ works as long as the culture can actually grow at $D$. Above some dilution rate it cannot — cells are being flushed out faster than any amount of substrate could replace them — and the steady state simply ceases to exist. That threshold is the critical dilution rate, and $D_{crit} = \mu_{max} S_0/(K_s + S_0)$.

The form is Monod's, evaluated at the FEED concentration $S_0$ rather than at the residual concentration inside the vessel. That choice trips people, and the reason for it is worth following. At the instant of washout there are effectively no cells left to consume anything, so the vessel contains the feed medium essentially unchanged; the fastest the last surviving cells could possibly grow is the rate that concentration supports. Using the residual substrate here instead gives a critical dilution rate that looks alarmingly low and is simply wrong.

On any normal fermentation medium the distinction between $D_{crit}$ and $\mu_{max}$ barely matters. A feed charged at 10 g/L against a $K_s$ of 10 mg/L puts $S_0$ a thousand times above $K_s$, and $D_{crit}$ comes out at 99.9% of $\mu_{max}$. The rule of thumb "washout happens at $\mu_{max}$" is serviceable, and it is the version most people remember. It stops being serviceable when the feed is dilute by nature, and that is not a laboratory curiosity — it is the wastewater case. A plant treating weak influent has $S_0$ comparable to $K_s$, $D_{crit}$ falls well below $\mu_{max}$, and the process cannot simply be run faster because the organisms have nothing to grow on. The hydraulic residence time has to be lengthened rather than shortened, or the solids have to be retained separately from the water, which is what settling and sludge return accomplish.

Washout is not gentle and it is not recoverable. Approaching $D_{crit}$ the steady-state biomass falls away steeply while the residual substrate climbs to meet $S_0$; the vessel is producing less and less biomass from more and more feed. Cross the threshold and the remaining cells are flushed out exponentially at rate $(D - \mu)$, which for a normal chemostat is a matter of hours. There is no recovery short of stopping the feed, letting the survivors grow back in batch, and starting again — or reinoculating outright. A long, expensive continuous run ends in an afternoon.

Because of that, prudent practice runs well below the calculated threshold, commonly at 60 to 80 percent of $D_{crit}$. The margin is not timidity. $\mu_{max}$ drifts with temperature, with pH, with the batch of medium components, with foaming and antifoam additions, and over a long run with the strain's own evolution — a chemostat is a selection device, and it selects for whatever grows best under the conditions imposed, which is not always the phenotype you started with. A vessel run at 95% of its calculated washout point is a vessel one bad thermostat away from an empty tank.

There is a related quantity worth knowing about even though it is not on this page: the dilution rate for maximum BIOMASS PRODUCTIVITY, $DX$, which is not $D_{crit}$ and is not close to it. Biomass output rises with $D$ at first, since you are making cells faster, then collapses as $D$ approaches washout and $X$ itself goes to zero. The maximum sits somewhere below $D_{crit}$, typically around 80 to 95 percent of it depending on $K_s/S_0$, and it is the operating point a production chemostat is actually aimed at.