Analysis of N2, O2, CO2, H2O and H2 in Ar and He

[JI2002]

We want to analyze 1-10 ppmv level of nitrogen, oxygen, carbon dioxide, water and hydrogen in 75% argon and 25% helium mix. The only method I'm familiar with is to use GC/TCD, but it's not sensitive enough to see down to 1 ppmv level. Does anyone know of a method that can be used for this analysis. Also I'm thinking how the gas manufactures test the impurities in their high purity (99.9999%) gas cylinder products? Thanks for any input.

[Bruce Hamilton]

I'm sure somebody from Matheson, Air Liquide or BOC would be happy to describe their methods. You should ask their technical people.

My recollection is that they use GC-TCD for macro components, FTIR with a very long pathlength gas cell ( IR signal is reflected back and forth through the gas about 20 times ) for some minor components , special detectors for other minor components ( eg inline paramagnetic for oxygen, inline electrochemical moisture meter for water ) , GC-FID with cryotrapping for trace HCs. etc. etc.

http://www.lne.fr/metrologie_francaise/ ... trogen.pdf

Bruce Hamilton

[JI2002]

Thanks for the reply. The information is very helpful.

[AICMM]

The impurities listed are a piece of cake in Helium. Just use an HID like mine or Valco's. The oxygen in argon is much harder to do since they normally co-elute but it is possible to get them to separate in which case an HID will work as well. Most air sep plants support an O2 analyzer that would take care of this problem and then you can do everything else with HID.

Write me if you need more info. aicmm at flash.net

Best regards.

[JI2002]

From the information I've gathered so far, there are several techniques available for the analysis, but none can do it all alone:

GC/TCD: Ar and He
FTIR: CO2, CO, CH4
Moisture analyzer: H2O
Oxygen analyzer: O2
GC/DID: Ar, CH4, CO, H2, N2, O2, CO2(with different columns and different oven temperature programs)

AICMM: I found in the following website that DID has some advantage over HID because it doesn't need radioactive source. Can you comment on that?

http://en.wikipedia.org/wiki/Discharge_ ... n_detector

[AICMM]

Three DID's that I am aware of. Gow-Mac, Valco, and mine are all non-radioactive and use various forms of discharge to excite helium and yield metastables. These metastables then ionize the analytes of interest. All three require leak tight systems since they will easily ionize any air leaking into the system. Commenting beyond that is going to become a sales pitch so I won't go there on this forum.

To my knowledge (and I watch pretty closely) nobody still makes a radioactive HID although Valco used to. Someone I know who still has one (as well as one of mine) says it is still the best way to detect hydrogen but I am not sure why.

If you want to measure the helium in your stream, probably the best way to do that is to use a TCD channel in nitrogen but it sounds more like a balance issue rather than a measurement issue.

Finally, it is reasonably easy to measure low levels of water with an HID in pure gases, just use a thick film non-polar column. The really tricky one in your list is the O2 in argon.

[AICMM]

A correction if I may. Just talked to the guy who has the radioactive HID and he said it is worse for hydrogen than the non-radioactive HID that he has. I am not sure where I had the reverse impression but that has been corrected.

Best regards.

[JI2002]

What's the best way to make low ppmv standard for N2 and O2? With N2 and O2 everywhere in the 80% and 20% level, how to prepare a low concentration standard contamination free? Thanks.

[Bruce Hamilton]

I've been out of the game a while, but the most cost-effective used to be aerosol cans of standards from Scott ...
http://www.scottsemicon.com/mixes/moxygen.html

If you're doing it routinely, the BOC and most other gas supply companies will make up certified mixtures in larger cylinders.

Very difficult to make really accurate standards, but if you only want an indicative value, I've made dilute standards using two syringes with luer fittings, one 1500 ml ( from Alltech I think ) for balance gas and one ordinary glass gas-tight 1 - 100ml syringe ( SGE ) for ambient air ( or Zero gas -if you have it ) , coupled together using a medical plastic three way valve.

After mixing the two gases, you can then dump the contents into a 40 litre 5-layer plastic gas sampling bag and fill with known volume of the main balance gas.
Because of the different densities, it's really important to thoroughly mix the contents at each stage.

You can then repeat the exercise with an aliquot from the first mixture to obtain trace values. There are also exponential gas dilution flasks, but I've never had much luck with them.

If it's important to obtain accurate values , I'd buy the certified standards, but the above method is good enough for developing methods.

Please keep having fun,

Bruce Hamilton