Collected from eight mesocosms, poured into to eight 4-liter-bottles, distributed into 96 tiny glass bottles. And then incubated in a pool with running water for the next 24 hours. That’s the way my samples take before being measured for the amount of oxygen produced or consumed in a day. Until this point you don’t really get to see anything, since it’s mainly pouring water from one container to another. But when the incubations end, the addition of the chemical reagents makes for some nice reactions. Manganese chloride and potassium iodide turn the crystal clear water sample into a glass snowball filled with light-brown flakes that slowly sink to the bottom. After that you add some sulfuric acid, which transforms the snowy precipitate into some sort of half solid brown jelly. Once you put the bottle on a stirring plate though, the jelly dissolves completely and gives the whole sample a nice amber color.
After all this we can start the actual measurement. In order to know how much oxygen was dissolved inside the sample, the DOA (Dissolved Oxygen Analyzer) starts adding small amounts of sodium thiosulfate. The more you put into the sample, the clearer the water gets. Until it once again reaches full transparency. The mechanism behind this reaction is actually quite simple: Manganese chloride is oxidized by the oxygen dissolved in the water. After adding some acid the manganese is reduced by the iodide, which in turn is left in the sample as iodine. The amount of iodine produced this way equals double the amount of oxygen that was initially in the sample and can then be quantified through the titration with thiosulfate.
This technique dates back as far as 1888, when it was developed by the Hungarian chemist Lajos Winkler. In the era of oxygen sensors it may be seem a bit cumbersome, but is very accurate and has become less labor-intensive thanks to automated sample analysis and modern computers. Repeat this process “only” 96 times while incubating the bottles under different conditions (total darkness, daylight and no incubation time at all) and you can tell how much oxygen the organisms in the sample are actually respiring and producing per hour! And that’s what we want to know.
1 – Bottle ready for the incubation
2 – The “Snowball” e.g. Manganese chloride and Potassium Iodide within the sample
3 – Acid addition turning the precipitate into brown jelly
4 – Final color of the sample after stirring and before being measured