In Oxygen at the start the yeast took a full dose of oxygen within an hour of pitching. We left the probe in the beer for the rest of batch #24 – 30 days, from pitching to packaging – to see what happens once the yeast has done its job.
In short:
- During active fermentation the reading stayed at the sensor’s floor. The yeast takes whatever gets in.
- Once fermentation slowed, the oxygen began to creep up, and each step down in temperature lifted it further.
- In cold storage it rose from 0.15 to 0.21 mg/L – about 5 mg of oxygen in 75 litres of beer.
- Most of the oxygen in the finished beer came in after the yeast had stopped working.

Active fermentation
For the first five days, while the gravity fell from 1.064 to about 1.019, the reading stayed at 0.036–0.046 mg/L. That is the sensor’s floor: it reads about 0.034 mg/L when there is no oxygen to measure. The first dry hop addition, a day and a half in, didn’t show at all – the yeast was busy and took whatever came in with the hops.
Warm conditioning
With the fermentation done, the beer stayed at about 22 °C for another week and then a few days at 24 °C. The oxygen crept up slowly, from 0.042 to about 0.06 mg/L. The yeast was still there, but no longer using much.
Soft crash and cold crash
Each step down in temperature lifted the reading. The soft crash to 14.7 °C took it to about 0.08 mg/L, and the cold crash to 0.5 °C to 0.14–0.15. The second dry hop addition went in during the soft crash and the third during the cold crash. Part of each step is simply that cold beer holds more oxygen at the same partial pressure, but the partial pressure went up too, from about 1 mbar during the fermentation to 2.2 mbar after the cold crash – so some oxygen really did come in.
Cold storage

In cold storage the reading was flat for three days. Then it rose in steps over the next four days, from 0.15 to 0.21 mg/L, and stayed there until packaging. In 75 litres that is about 5 mg of oxygen. Averaged over the nine days it’s about 25 µg an hour – the same order as the gasket tests found for a fermenter’s gaskets and fittings. But it didn’t come in evenly: the steps look more like single events, like connecting to a gas line that has stood closed.
Our notes in Grafana show what we did. We dumped trub several times – twice during the soft crash, once at the cold crash and once more early in the cold storage – and none of those stood out. The rise started with the dump and tasting on day 21. The dump lowered the pressure, and the fermenter took about 25 g of CO₂ to get back up. From then on the oxygen rose for four days, faster than the cold storage before it. The CO₂ is not the likely source: we topped up the pressure three times earlier in the cold storage, with 4–11 g of CO₂ each time, and the oxygen didn’t move. Nor, probably, is the dump itself – we purge the whole pipe from the dump valve before dumping, so no air should get in that way, and the earlier dumps didn’t show. Our main suspect is the sample valve. Not the sampling itself, but that the valve ends up in a position after a tasting where it doesn’t quite seal – as if it hadn’t been closed properly. The steepest rise came right after the dump and tasting, and after the next tasting a day later it slowed to about a third. Half a day later it picked up again without anything we know of, and from day 24 it rose in short bursts every afternoon – that pattern we can’t explain yet.
What it means
When the beer was packaged, it held about 0.21 mg/L – about 0.18 mg/L above the floor, or 180 ppb. The yeast had cleaned up everything until the fermentation was over, so practically all of it came in during conditioning, the crash and the cold storage.
We don’t have a plan for what to change yet. First we want to find out where the oxygen in batch #24 came from.