Oxygen Demand
Bacteria eating waste breathe the river's oxygen, and BOD measures their appetite.
A milk tanker overturns on a bridge, and its load drains into the creek below. Milk isn't poison. You'd drink it. Yet within days, fish are floating belly-up for a long way downstream. Nothing in the milk killed them. They suffocated.
Here's how food takes the oxygen out of a river, and how a glass bottle and five days of patience put a number on the danger.
How food suffocates a river
Water holds remarkably little oxygen. Clean water at 20°C, saturated with air, holds about 9 mg/L of dissolved oxygen, or DO. That's nine parts per million, and everything in the river that breathes shares it.
Now add food. Bacteria in the water eat organic matter, and they breathe while they do it, pulling DO out of the water in proportion to what they eat. Milk counts as food. So does sewage: typical municipal sewage holds enough to use about 200 mg/L of oxygen, more than twenty times what the water holds at any moment. When the bacteria breathe faster than air can replace the oxygen, DO falls, and the fish are the first to feel it.
Biochemical oxygen demand, or BOD, puts a number on this: the oxygen, in mg/L, that bacteria use while eating the organic matter in a sample. It's really a measure of food, read out in oxygen.
Bacteria don't eat all at once, so a BOD needs a clock attached. counts the oxygen they use in the first five days. Ultimate BOD, written , counts what they'd use given all the time they want. is only part of , and BOD Kinetics connects the two.
Three ways to count the same food
BOD is one of three numbers for oxygen demand, all in mg/L of oxygen. They differ in who, or what, does the eating.
- BOD is what bacteria will eat. You give them time and measure the oxygen they use.
- COD, chemical oxygen demand, is what a strong chemical oxidant can burn. A lab heats the sample with the oxidant, usually potassium dichromate in hot acid, and converts the amount used into the equivalent oxygen. No bacteria are involved, and it takes hours instead of days.
- ThOD, theoretical oxygen demand, is what a balanced equation says complete oxidation would take. It isn't a test at all. You work it out on paper, which takes knowing exactly which chemicals are in the water, so it's mostly for pure compounds.
For the same wastewater, COD normally comes in above BOD, often about twice for municipal sewage. Hot dichromate isn't a picky eater. It burns nearly everything bacteria can digest, plus plenty they can't, such as the lignin in wood fibers. And it isn't on a five-day clock.
The Handbook won't help you here. It mentions COD only in passing and prints nothing on ThOD.
The bottle test
The test is almost disappointingly simple. Put some of the waste in a bottle, measure its DO, seal it, wait, and measure again. The oxygen that disappeared is what the bacteria breathed. The conditions keep that last sentence true and make one lab's number comparable with another's:
- Sealed, so no air can put oxygen back.
- In the dark, so algae can't photosynthesize and put oxygen back either.
- At 20°C, because bacteria eat faster when they're warm, and every lab has to measure the same thing.
- For five days, which is simply convention. The usual story credits a British royal commission around 1900, reasoning that few British rivers take longer than that to reach the sea.
There's one catch. Fill the bottle with raw sewage and the bacteria burn through its 9 mg/L long before day five, and all you learn is "more than 9." So you use a little sample and fill the rest with dilution water: clean water saturated with oxygen, with nutrients and a pH buffer for the bacteria but no organic food. The standard BOD bottle holds 300 mL. Say you put in 6 mL of sewage and top it up with 294 mL of dilution water.
The bottle lost . But the sewage is only of what's in the bottle, and dilution water brings no food, so that one-fiftieth did all the eating. Full-strength sewage would demand fifty times as much: 210 mg/L. Dividing the drop by the sample's fraction scales it back up, and that's the formula in the Handbook's BOD Test Solution and Seeding Procedures section:
Here is the DO right after preparation, is the DO after five days at 20°C, and is the sample's volume divided by the bottle's volume, as a decimal. The Handbook labels the left side only "BOD," but its is a five-day reading, so this is . It doesn't mention the 300 mL bottle or the dark, so those are yours to remember.
When the sample has no bacteria
All of that assumes there are bacteria in the bottle. Raw sewage is crawling with them. A chlorinated effluent may have almost none left, and some industrial wastes kill them off. Bottle those as they are and the DO barely moves, food or no food. So the lab adds seed, a small dose of bacteria, usually stirred into the dilution water.
Seed fixes one problem and creates another. The seed bacteria bring a little food of their own and breathe while they eat it, so part of the bottle's drop isn't the sample's doing. To find out how much, the lab runs a second bottle alongside, the seed control: seeded dilution water and no sample. Its drop is what the seed costs when it has a full bottle to itself.
The sample bottle got a little less seed than the control, because the sample took up some of its room. It holds 270 mL of seeded dilution water against the control's 300 mL, so it got of the control's seed, and the seed's share of its drop is . Take that off and scale up as before, with : . The Handbook's BOD Test Solution and Seeding Procedures section writes it this way:
and are the seed control's DO before and after incubation. The ratio compares the seed in the sample bottle to the seed in the control, in that order, because you're scaling the control's drop to fit the sample bottle. The Handbook states it as volumes of seeded dilution water. And still counts only the sample.
Which bottles can you believe?
A lab rarely knows a sample's strength beforehand, so it sets up several bottles at different dilutions and asks two questions of each one:
- Did the DO drop by at least 2 mg/L? A smaller drop is too close to the meter's error, and dividing by magnifies that error.
- Is at least 1 mg/L left at the end? If not, the bacteria may have run short of oxygen before day five, and the drop understates the demand.
Those two rules leave a window. With too little sample the drop is too small, and with too much the oxygen runs out. The Handbook prints neither rule; they come from the standard lab procedure.
One more thing can spoil a reading. The BOD so far is all carbonaceous demand: bacteria eating carbon compounds. A different group, the nitrifying bacteria, eat ammonia, breathing the same DO while they turn it into nitrate. That's nitrogenous demand. Nitrifiers grow slowly, so in raw sewage their demand usually doesn't show until after the five-day test is over. But a well-treated effluent can already be full of them, and then they start in on ammonia inside the five days and inflate the result. When a lab wants carbonaceous demand alone, it adds a nitrification inhibitor. The Handbook prints nothing on any of this.