Steel tanks absorb O2?

[oceancurrent]

I was doing my mid-winter gear self-servicing today and noticed a rather interesting anomaly. My two steel tanks (PST E-7 HP 80) registered 0.5% lower O2 content than my two aluminum tanks (Luxfer). I repeated the analysis 5 times comparing the aluminum and steel tanks to establish accuracy. The measurements were consistent (0.5% less O2 in the steel tanks). All tanks were filled at the same store about 3 months ago in the same week. Furthermore, they have all been kept at 1000 psi for the period of storage.

I've never filled the aluminum tanks with Nitrox, so we can eliminate the suspicion of having residual Nitrox in the tanks at the time of the fill. Assuming that the dive shop didn't accidentally add trace amounts of "free" Nitrox to the aluminum tanks at the time of the fill, it may be interesting to consider the following hypothesis:

Either:

A. The steel tanks use up some O2 during long term storage. Perhaps there is an ongoing rusting process on a micro scale? Or O2 diffuses through the steel during long term storage?

B. A gas other than O2 gets released in the steel tanks through some chemical/physical process.

C. The O2 concentration drifts in the aluminum tanks.

D. All of the above.

E. None of the above.

While it would be impossible to prove any of the above without setting a controlled experiement, I thought it would be interesting to ask the chemists/physicists amongst you for an opinion on this. Do you think that this "O2 shift" in Al vs. Steel tanks in 3+ month storage is possible? Any idea what may be the explanation if this is the case? Has anybody seen this before?

 

[Rick Murchison]

I was doing my mid-winter gear self-servicing today and noticed a rather interesting anomaly. My two steel tanks (PST E-7 HP 80) registered 0.5% lower O2 content than my two aluminum tanks (Luxfer). I repeated the analysis 5 times comparing the aluminum and steel tanks to establish accuracy. The measurements were consistent (0.5% less O2 in the steel tanks). All tanks were filled at the same store about 3 months ago in the same week. Furthermore, they have all been kept at 1000 psi for the period of storage.

I've never filled the aluminum tanks with Nitrox, so we can eliminate the suspicion of having residual Nitrox in the tanks at the time of the fill. Assuming that the dive shop didn't accidentally add trace amounts of "free" Nitrox to the aluminum tanks at the time of the fill, it may be interesting to consider the following hypothesis:

Either:

A. The steel tanks use up some O2 during long term storage. Perhaps there is an ongoing rusting process on a micro scale? Or O2 diffuses through the steel during long term storage?

B. A gas other than O2 gets released in the steel tanks through some chemical/physical process.

C. The O2 concentration drifts in the aluminum tanks.

D. All of the above.

E. None of the above.

While it would be impossible to prove any of the above without setting a controlled experiement, I thought it would be interesting to ask the chemists/physicists amongst you for an opinion on this. Do you think that this "O2 shift" in Al vs. Steel tanks in 3+ month storage is possible? Any idea what may be the explanation if this is the case? Has anybody seen this before?

If - pretty big if - your tanks all actually had the same % oxygen in 'em to begin with then you would be wise to drain and inspect those steel tanks. A .5% decrease in oxygen content at 1000 psi would represent substantial rusting. (a little over a half ounce of steel consumed)
However... my bet would be that there was some gas in the tanks when they were filled that resulted in the mismatch, or the difference was in the gas at the shop (did they run the compressor during or between fills? A 0.5% difference isn't unusual from one run to another, or even during a run).
Rick

 

[KOMPRESSOR]

If - pretty big if - your tanks all actually had the same % oxygen in 'em to begin with then you would be wise to drain and inspect those steel tanks. A .5% decrease in oxygen content at 1000 psi would represent substantial rusting. (a little over a half ounce of steel consumed)
However... my bet would be that there was some gas in the tanks when they were filled that resulted in the mismatch, or the difference was in the gas at the shop (did they run the compressor during or between fills? A 0.5% difference isn't unusual from one run to another, or even during a run).
Rick

I agree with Rick. 0,5% difference is a pretty good match on two fills, actually. Yes, rust would steel some O2, but you have no indication of decrease in O2 level here. I wouldn't worry. You know yourself how your tanks have been filled, and the eventual rust will show up at inspection time.

As long as the air you get is dry, your steel tanks will last "forever" in your diving career.

 

[Orlando Eric]

Just so that I am on the same page here. We are talking one half of one percent. (.05%). I do not know what sort of analyzer you are using but I seem to remember that the "sniffer" types have a margin of error greater than that by specifications.

If that is one of those nagging things that is going to crawl around in your head when diving then get a visual.

 

[KOMPRESSOR]

Just so that I am on the same page here. We are talking one half of one percent. (.05%). I do not know what sort of analyzer you are using but I seem to remember that the "sniffer" types have a margin of error greater than that by specifications.

If that is one of those nagging things that is going to crawl around in your head when diving then get a visual.

0,5% exactly is what I read here. Not 0,05%, which would be unmeasureable with most analyser I know of!

 

[DA Aquamaster]

The sensors in most analysers are accurate to about +- 1%. So .5% is lost in the noise of the sensor's margin of error and is nothing I would worry about.

On the other hand, oxidation in a steel tank does, obviously, use up oxygen, so it coud indicate some rusting going on and it's something you might want to follow up on at your next VIP. Tumbling could be in your tanks immediate future.

 

[divermatt]

I'm not an expert in steel vs. AL, or a chemist/metalurgist. How about we do this: Get a separate O2 analyzer in there. You seem rather a scientific sort, but you only have used one analyzer, which makes it a variable. Also, note the tolerance +/- % for each reader. Oh, and use a tank of air between each reading to baseline the sensor. Without exploring your exact methodology (did you read both steels, then read both ALs????) I'm curious to see what you find. I cannot imagine 2 steel tanks oxidizing .5%. If you have any feeling that this was the case, it's time for a Vis/hydro.

 

[KOMPRESSOR]

I'm not an expert in steel vs. AL, or a chemist/metalurgist. How about we do this: Get a separate O2 analyzer in there. You seem rather a scientific sort, but you only have used one analyzer, which makes it a variable. Also, note the tolerance +/- % for each reader. Oh, and use a tank of air between each reading to baseline the sensor. Without exploring your exact methodology (did you read both steels, then read both ALs????) I'm curious to see what you find. I cannot imagine 2 steel tanks oxidizing .5%. If you have any feeling that this was the case, it's time for a Vis/hydro.

No reason to! In this case 0,5% isn't justified as a "drop" to begin with, as the owner don't know exactly what the O2 level was just after filling. There may be differences between the two fills and there may have been a difference in O2 level in the two tanks BEFORE the fills. Therefore, don't bother.

 

[DA Aquamaster]

No reason to! In this case 0,5% isn't justified as a "drop" to begin with, as the owner don't know exactly what the O2 level was just after filling. There may be differences between the two fills and there may have been a difference in O2 level in the two tanks BEFORE the fills. Therefore, don't bother.

I basically agree with you - but not entirely.

If we were talking about a nitrox fill then it would be very clear that a diver would not have any idea precisely what went into the tank originally unless it were marked exactly on the contents lable. Even then, using a different sensor and/or a different analyzer or even the same sensor after a few more months of use or even shelf life exposed to air could cause a .5% difference. In that case the only way to really be able to tell if oxidation is occurring is to compare the known when filled contents of an AL tank with the known when filled contents of a steel tank. Then a larger difference between the contents label and contents of the steel versus AL tanks now could indicate some oxidation occurring. The AL tank would serve as the control and would eliminate any random error caused by using a different sensor, analyzer etc.

Of course...what worries me here is that (the way I read the post) they were all air fills, not nitrox fills. I cannot remember the last time I saw an air sample that was anything other than 20.9%-21% as determined by a consistent comparision of a free atmosphere calibration with the tank contents. I always get a 20.9% reading from a tank of air using a 20.6% open air calibration. The .3% difference is consistent and due to the slightly higher pressure inside the sample chamber attached to the sensor on my flow restrictor equipped analyzer.

So assuming the compressor was not "burning" O2 in some strange failure, you can reasonably expect that all the tanks, even if filled a week apart, would be filled with the same 20.9 percent O2 mix. That would then create the same controlled test as outlined above and a .5% decrease in the contents of the steel versus Al tanks could be significant asuming the sensor readings are stable and consistent.

To be sure, I would calibrate the sensor off a third source and then measure any relative differences using exactly the same testing procedures.

There are still other possible testing variables. If you are using an analyzer with a restrictor and enclosed sensor chamber, the results are going to be very consistent. If on the other hand you are just blowing the sample across a sensor held up to the valve in one of those little Tee thingies, differences in valves and flow rates from the tank could cause the .5% discrepency.

 

[oceancurrent]

Thank you all for the interesting comments!

I am not worried about the tanks at all - they were purchased new just half a year ago. I would be very surprised if I find any rust in them on the next visual inspection. However, I will err on the side of cautioun and get them looked into. I was just curious to see if anybody has any interesting thoughts to share on the subject of gases stored under pressure in metal containers.

The analyzer I use is RC Dive Technologies with an Analytical Industries sensor (PSR-11-39). It has proven accurate within 0.1% when comparing two mixes side-by-side in a stable environment using a 2 lpm flow restrictor (which was the case here). I calibrated the analyzer against open air at 20.5 (in my experience 0.4% compensates perfectly for the pressure difference between open-air and flow-restrictor measurements). The sequence, in which I measured the tanks, was Al1-Al2-St1-St2-Al1-St1-Al2-St2-Al1-St2-Al2-St1. The procedure was done in a few minutes without any interruption or recalibration. The analyzer gave a series of stable readings: 20.9-20.9-20.4-20.4-20.9-20.4-20.9-20.4-20.9-20.4-20.9-20.4. Thus, I am confident that there is 0.5% (+/- 0.1%) difference in the O2 content of my aluminum and steel tanks.

I agree that more than likely the tanks had slightly different FO2 after the fills for one reason or another, which would explain everything. However, if they didn't (which is not entirely improbable), the next best explanation would be the type of container and chemistry. The possibility that the O2 content may shift noticeably over time depending on the type of container (oxydation, diffusion?) is quite intriguing.

You know what... Since I've got three more months of tank storage left before the dive season, I might as well set up a controlled experiment to satisfy my curiosity. I think I will bring the tanks at the dive store, drain them, look into the steel tanks for rust, fill all the tanks to 1000psi with air from the same source, let the their temperature equalize with the environment at home for a few hours, perform a reference analysis, and then let them "age" for three months. This could turn out interesting.