What COD fractionation is, and why a design without one is a design for somebody else's water
Two waters can arrive at the same total COD and need different plants. The number that separates them is not on a standard laboratory report.
A discharge monitoring report gives you total COD. It is one number, it is cheap, and every plant has years of it. It is also almost useless as a design input on its own, because it says how much oxygen demand arrived and nothing at all about how fast the biology can get at it.
Fractionation is the step that splits that single number into the parts an activated sludge model can use. It is the difference between designing for your water and designing for a category your water happens to belong to.
The four fractions
Total COD divides on two axes at once: soluble or particulate, and biodegradable or inert. That gives four fractions, and each one behaves differently in a reactor.
| Fraction | What it is | What it does in the plant |
|---|---|---|
| Readily biodegradable | Soluble, small molecules — volatile fatty acids, sugars, alcohols. Taken up in minutes. | Sets the peak oxygen demand and drives denitrification. It is the carbon an anoxic zone actually has available. |
| Slowly biodegradable | Mostly particulate and colloidal. Must be hydrolysed before it can be taken up, which takes hours. | Sets the reactor volume you need. Hydrolysis, not uptake, is the rate-limiting step. |
| Soluble inert | Passes through untouched. | Appears in your effluent COD at essentially the influent concentration. No reactor volume removes it. If your permit is on COD, this fraction alone can decide whether the permit is achievable by biology at all. |
| Particulate inert | Not degraded, but captured in the floc. | Accumulates in the mixed liquor and leaves as sludge. It raises MLSS and sludge production without contributing any treatment capacity. |
Why the split matters more than the total
Consider two industrial effluents that both report the same total COD on the same laboratory sheet. One is a brewery-type stream that is largely readily biodegradable; the other carries a large soluble inert fraction from a process using a recalcitrant solvent.
Designed from the total alone, they get the same plant. In reality:
- The readily biodegradable stream needs less reactor volume than the total suggests, because uptake is fast — but it needs more installed aeration capacity at the peak, because that demand arrives all at once.
- The stream with the large inert fraction needs no more volume for the inert part, because nothing happens to it — but no amount of volume will get its effluent COD below the soluble inert concentration. If the permit sits below that number, the answer is not a bigger biological plant. It is source segregation, or a different process entirely.
That second case is the expensive one to get wrong, and it is not visible in the total. A plant gets built, commissioned, and never meets its permit — and the post-mortem finds that it could not have, from the first drawing.
The one-sentence version
The total COD tells you how much oxygen demand arrived. The fractionation tells you which of it your plant can do anything about.
How the fractions are measured
The measurements are ordinary, and any competent laboratory can do them. What is not ordinary is asking for them.
The soluble/particulate split
Total COD on the raw sample, and COD again on a filtered sample. The filtration is where the care is: a standard glass-fibre filter passes colloidal material that behaves as particulate in the reactor, so a flocculation-and-filtration step before the measurement is what separates truly soluble from operationally-soluble. Skip that and the readily biodegradable fraction reads high.
The inert fractions
The soluble inert fraction is obtained from a long-duration batch test: aerate a sample until nothing further is degraded, and the soluble COD remaining is the soluble inert. It is slow — days — and it is the single most valuable number on the sheet when the permit is on COD.
The particulate inert fraction is normally back-calculated from observed sludge production over a period in which the load and the wasting rate are known.
What respirometry adds
A respirometric test measures oxygen uptake rate against time on a real sample of your water and your sludge. The shape of that curve separates the readily biodegradable fraction from the slowly biodegradable one directly, rather than by inference: fast uptake falls away as the readily biodegradable material is exhausted, leaving the slower hydrolysis-limited plateau.
It also does something a composition analysis cannot: it shows inhibition. If your water contains something that suppresses the biology, respirometry is where it appears — as a curve that does not rise the way the COD says it should. On an industrial effluent that is a common and expensive surprise, and it is far better found in a flask than in a commissioned reactor.
The same argument applies to nitrogen
Nitrogen splits too — ammonium, soluble organic nitrogen, particulate organic nitrogen, and nitrate — and the split matters for the same reason. It is also where model defaults do the most damage, because default composition factors are calibrated on municipal sewage and will happily manufacture nitrogen that a strong, nitrogen-poor industrial water does not contain.
We know that one precisely, because it happened in our own tool. On our worked example the default composition factors delivered 114.6 mg/L of TKN against the 60 mg/L that had been entered — a 91% overstatement. The measured TKN now sets the total and the composition factors distribute it.
The full verification is published, including that item and the eleven others it found.
What to ask for
If you are commissioning a characterisation, or judging one somebody else did, this is the list:
- Total and soluble COD, with the filtration method stated.
- BOD₅, TSS and VSS.
- TKN, ammonium, nitrate, total phosphorus, alkalinity, pH and temperature.
- The four COD fractions, with the method used for each.
- Respirometry where the water is industrial, or where inhibition is plausible.
- Sampling across the real production cycle — including the days that are not typical, because those are the days the plant fails.
And one question to ask of any design put in front of you: what fractionation is this based on, and where did the numbers come from? If the answer is a category — "typical brewery wastewater" — the design is for a typical brewery.