DO sampling techniques

[Mike Elliott]

Hey all,

Just got a Mettler-Toledo optical DO meter and am wondering about SOP for testing at various points (e.g. pre-filter, post-filter, brite, and package.)

Main question is: how long does it take for levels to drop in cold beer? Do we need to sample cans immediately off the line? Does anyone have any literature on reaction rates, oxidation products, etc?

I'm asking because we can pull a can from the line and get a peak reading of around 80 ppm, but that drops to 10-15 in less than a minute. Is that a realistic rate of oxidation?

Any reading suggestions would be appreciated.

Cheers,

Mike
Kettlehouse Brewing Co.
Missoula MT

[Hazell]

I have not used one of those meters myself. We use Orbisphere units.

However, if they are like ours, we ignore the peak reading and wait until the DO reading is stable. In your case I would be looking at the 10-15 reading. Are you sure the units are not in ppb? 10-15ppm is a VERY high DO.

[Mike Elliott]

Sorry, definitely ppb. As far as I can tell, we don't run our beer over an O2 stone before packaging.... too much excitement for my typing fingers to follow.

[KWLSD]

Mike,

You may find that the initial 'peak' reading is being caused by beer mixing with air in the inlet line of the meter (or water if you have flushed it though on a tap after previous use). With air at ~20% O2 it takes only a very small amount to add 60-70ppb to a liquid when mixed.

Alternatively it could be interference due to microbubbles of CO2 created during filling or any temperature compensation in the meter taking time to stabilise.

Always wait for a stable reading.

It is very unlikely you will be getting oxidation fast enough for the levels to drop as reported due to that phenomenon.

Regarding sample points: The ones you list are good for final product; however it would be useful data for you to do some measurements from fermenter to maturation, to build up a picture of your process at that stage.

Theoretically at the end of fermentation you should have zero DO2, but if you're pumping, etc. then it would be possible for a leak to introduce Oxygen into the beer stream, which would then cause flavour damage during maturation.

My view has always been that it's better to know what a 'good' baseline looks like - then if anything changes, you have a better idea of where to start looking!

[kai]

I agree, if the reading is dropping from 80 to 10-15 from the same can, hooked up to the meter then it's interference from the meter inlet not being fully purged, or response time from the meter itself (or are optical units much faster than membrane-based probes?).

If you are pulling a can off the line immediately and getting a reading of 80ppb, then letting a different can sit for 15 minutes and getting a reading of 10-15, then I too would be wondering where those rascally O's have gone. My gut says the beer's not going to oxidise the difference that quickly, but I really wouldn't say for sure.

PS Mr Hazell, I look forward to you buying me a beer next week

[BrianWiersema]

You need to bring your meter down to around the expected DO and temp levels before taking a reading. I always run some "meter prep/sacrificial throw away" beer through before finding a stable reading. This means that when I check at the brite tank, or before after pumps, etc. I let beer run through until it stabilizes. If I am checking bottles, I pull 2 full rounds off the filler, plus an extra 4 bottles. the 2 rounds go right into the fridge, and then I push the 4 bottles through the DO meter. Those "throwaway bottles" bring the DO down to the approximate DO and temp (very important) I will expect in my analysis bottles. Thus the lag time for the DO meter is minimal.

You should analyze your packaged product immediately after it is filled to get the DO content at that time. I take 2 rounds: one that I do not shake the bottle, and one that I shake the bottle right before taking the DO. The unshaken one gives me the DO of the beer in the bottle, and the shaken gives me the DO in the beer + DO in the headspace so that I can do TPO calcs.

hope that helps. Oh, and I would love to get some of your beer out here in CA! delicious!

[KWLSD]

...are optical units much faster than membrane-based probes?.

The optical systems we have seen in use aren't any faster than EC sensors.

One of my concerns with certain optical sensors is that they have a very low saturation level:

Typically an LDO optical sensor that is good at low levels (sub-100ppb, say) will saturate out at 2ppm. After that it will just read 'out of range', which isn't helpful when an excellent guide to meter performance is to put is on clean, cold water direct from a rising main supply.

Okay it's only a guide, but that should read around 8-9ppm.

Unlike EC sensors, this means you can't use one meter for the whole process from wort oxygenation to finished beer.

[dick murton]

The optical units I have use (5 different suppliers) all responded within individual scans. We had them set up at 2 second scans, and all of them identified peaks due to faulty mechanical seals on a pump (with long suction run to the pump) which the conventional meter hardly noticed, for instance peaking for two or three readings only up at hundreds of ppb, then dropping back to say 15 ppb within another few readings. The conventional Orbispheres only picked up a spike to about 30 ppb.

However, as with all O2 meters, you need to purge the sample system out thoroughly. But again, we could get consistent DO2 readings in about 30 seconds off bright beer tanks, as against 2 minutes for EC units.

The optical units do not need to be attemperated due to small cahnges in temperatures. However, they will need the optical membranes replacing more regularly if active at high temperatures, so you need to make sure you don't run hot CIP past them whilst they are active. But all the ones I tried have a variable thermal cutoff capability, and some are suitable for a switching input from PLC

They didn't appear to be as sensitive to entrained gas bubbles as EC units.

We had to put offsets in them of 10 to 15 ppb when measuring directly ex FV to gain confidence that the levels we were seeing were consistent and repeatable, as often they went to zero. An un-offset but freshly calibrated EC unit alongside eventually dropped to say 5 ppb, so there was a high level of consistency between old and new technology.

I do agree with the comment about limited range, but I know one or two suppliers are looking to provide full range units, rather than having to supply two different instruments, one for bright beer type levels (0 - 2000 ppb) and another for wort oxygenation (0 - 30 ppm). But against that, we have to put in heavier membranes (I think simply for physical strength) when using EC probes in wort lines only. Hand helds are fine for all uses with a finer membrane.

[KWLSD]

Dick - you are quite correct that heavier-gauge membranes are standard for inline EC sensors to increase robustness; however I would argue that if someone is experiencing pressure fluctuations high enough to damage membranes then things like pump start-up, valve opening sequences, etc. should be part of the solution, not simply a change of DO2 instrument, as I have seen done.

I am also concerned about your observation of an EC sensor taking two minutes for stability; this is way too long, especially with a reasonably sensitive membrane, when a properly service unit should (and does) respond in under 30 seconds.

The benefit in that case - as I have suggested - is that you can get good low-end (<100ppb) response but still have full-scale of 15-20ppm; something which as yet is beyond the capability of an optical sensor.

Optical units are also susceptible to spot degradation from stray light as well as CIP & high temperature; and an EC unit can also have a temperature cut-off preset to shut down the sensor where hot CIP is used.

Your comments about entrained gas and offsets are interesting; all-in-all, I don't believe it is quite time to write-off EC sensors, as some of the marketing guys would have everyone think!

However, I am aware this is now getting rather 'off thread', so I'll leave it there

[dick murton]

"if someone is experiencing pressure fluctuations high enough to damage membranes then things like pump start-up, valve opening sequences, etc. should be part of the solution, not simply a change of DO2 instrument, as I have seen done." Agree totally, but with the best will in the world, due to vessel sizes, piperuns etc, it is not always to eliminate them completely - worst case is drawing a slight vacuum.

I have not come across the stray light problem - any particular source ?

Thanks

[KWLSD]

Dick - worst case I ever saw was pressure spikes capable of punching an EC membrane through the support ring holes, even with the sensor fitted with a protection cap

Regarding optical degradation due to stray light - any light source that has an emission near the wavelength of the excitation LED will cause fluorescence and, ultimately, shorted the life of the spot. The sort of thing that comes to mind is inline sight-glasses, especially where they are illuminated.

NB. This refers primarily to inline installations rather than portable meters.

[dick murton]

Thanks very much for the info about the stray light. We are likely to be buying some more optical units to fit into new kit, so I can make sure the sensors are away from the sightglasses - especially as they will be illuminated (on demand, short fixed time only)

Does stray light affect the accuracy of measurement as well, by "overloading" the sensor, or is it simply accelerated wear ? Oh, and which of the suppliers have let this little gem loose ? Thanks again

[KWLSD]

Dick - I would suspect that it's not been considered by the manufacturers...or if it has, they're not saying!

However it would seem logical that a constant light source would cause some degradation but possibly a lesser effect on accuracy as the background fluorescence would be relatively stable (although that, in itself, may result in an offset - all depends on exactly how the sensor is designed to detect); an occasional light, eg. illuminated sightglass, would be more likely to give fluctuating results.

[AnalysisAnnie]

LDO sensors use ruthenium complexes impregregnated into the dots. As KWLSD has stated the probe works by exciting the complexes with pulses of blue light from an LED and the ruthenium complexes emit red light for detection. If the spots are subject to stray light i.e. via sightglasses etc excitation can take place. If the spot is consistently exposed degradation is bound to occur as this excitation is due to the presence of light of the correct wavelength and has nothing to do with the presence of oxygen. Ultimately these spots do require protection from high temperatures, caustic and high oxygen levels or degradation will take place.

[kai]

Some interesting points here, thanks!