O2 analysis in Nitrogen

[mbeckwick]

I'm having trouble analysing O2 levels(1-10ppm) in our plant nitrogen. The GC is a 4900 Varian GC, with a 10M Molsieve 5 Column. I believe that the argon and oxygen peaks may not be separating.? I've tried the column at ambient temperature and 90C, with no luck. Any ideas?
Thanks.

[chromatographer1]

If you want to separate argon and oxygen with a capillary column MS5A you will require subambient temperatures or very good luck. Many MS5A columns will absorb ppm levels of oxygen as well, and do so inconsistently just to be a real pain in the analysis.

Oxygen and argon can be separated at room temperature with a special order packed column available from Supelco.

If you wish to see example chromatograms email me at my work address and I can send them to you.

best wishes,

Rodney George
Supelco
rodney.george@sial.com

[MikeD]

To add to what Chromatographer 1 said:

When we need to resolve argon and oxygen without subambient we use a 25 m x 0.32 mm MS5A column at 30 deg C. You have to take care not to degrade resolution with too large injection volume or split ratio too small. Otherwise you will not be able to duplicate any supplied test chromatogram.

[Bruce Hamilton]

Were Ar and O2 separating in the past?. If so, you may need to revisit the column conditioning temperatures. If not consider the column suggested above.

My experience has been as described above, can be very variable, but a well conditioned packed or PLOT MS5A column will separate them for a while, but eventually becomes saturated? and then loses resolution quickly. the first five - ten 0.5 ml samples each day would separate, then the peaks would merge, and stay that way until an overnight thermal conditioning of the column.

Restek used to claim their Plot MS5A column would consistently separate Ar and O2. Ask them for details.

Another technique is to chemically remove the oxygen from a sample and analyse again. Not recommended for process situation, but could be helpful if you want to investigate whether they are coeluting.

Please keep having fun,

Bruce Hamilton

[chromatographer1]

The ability of the sieve in plot columns to separate oxygen and argon at low levels is dependent upon the amount of sample placed on the column and the concentrations of the two gases.

Different PLOT columns from different vendors will not always perform in the same manner due to the thickness of the layer and the binding agent used to secure the sieve to the column.

Overconditioning will cause the oxygen to tail and the accuracy of its measurement to be dubious. Experience and care is necessary to properly condition a PLOT column. Once damaged it is not usually possible for the column to be repaired.

Even so, a sample with a large amount of oxygen or a large amount of argon may prove difficult to analyze with any PLOT column.

A packed column, carefully packed with molecular sieve that has been properly cleaned and conditioned with tolerate much larger samples and be able to separate larger fractions of argon and oxygen. If glass lined or fused silica coated steel tubing is chosen then low ppm amounts of oxygen can be measured with a high confidence and reproducibility. AT-Steel and Silcosteel are commercially available products that immediately come to mind.

Having tested many lots of molecular sieve from several commercial sources I have seen a wide variation in performance at low levels of oxygen content. This holds for PLOT and packed columns.

Ask for examples of argon and oxygen separation before you buy so you don't waste time or money.

Disclaimer: I work for Supelco and we do have columns that WILL separate argon and oxygen even at hgh percent levels and will elute oxygen at low ppm levels at ambient conditions and even higher. There may be other vendors who can as well.

(sorry for the pitch but Supelco helps customers who find the columns they purchased elsewhere often don't get the job done.)

Rod

[Bruce Hamilton]

I thought I had suggested that the OP consider Rod's column if they hadn't achieved separation before, but I was also trying to address the situation where the separation had previously worked.

I wasn't suggesting over-conditioning the column, but following the manufactuer's recommendations. I've used Chrompack MoleSieve 5A PLOT 10 um film and HP-Plot Molesieve columns for O2 in N2, with each having a different recommended conditioning regime, eg HP was seven hours at 200C, which worked for me, and the Chrompack could go to a higher temperature but I found 190C/Overnight was suitable for my samples. YMMV.

I mentioned Restek because, at the time I was doing trace Oxygen in Nitrogen, their column was claimed to be superior for Ar/O2 to the columns I was using.

I wasn't intending to suggest it was the best, or only, option, just that they might offer further insight into the issue. Looking at their www site, it doesn't seem much different to what I was seeing with other PLOT columns.
http://www.restek.com/aoi_petro_A009.asp

Sorry if anyone was annoyed, and hopefully nobody was offended...

Bruce Hamilton

[chromatographer1]

Thank you Bruce, for sharing your experience with different columns. I am sure this information will be useful to analysts who prefer capillary columns. I would entertain any additional information concerning other manufacturers of Mole sieve columns if anyone would share. I have heard reports from users that certain film thicknesses were unsatisfactory in giving adequate separation of argon and oxygen. Of course, this may only be a factor of individual columns failing to work, perhaps caused by their previous use or conditioning. I appreciate hearing more input from users and thank you again Bruce for your valuable contribution. I hope I did not sound annoyed (I am not by the way ) and only intended to share results I had received from users.



This is the separation of argon at 0.9% and oxygen at 21% in nitrogen.
The temperature is 27°C and the injection volume is 100µL using a HP 6890 µTCD @ 150°C

The column was 9 meters in length and the flow rate of helium carrier is 30cc/min.

best wishes,

Rod

[salwa_meredith]

Halo,

We are currently using: Agilent 6890; Molecular Sieve 5A; to analyze oxygen and nitrogen separation peaks.

From the chromatograms, we noticed both oxygen and nitrogen gases have exactly the same peaks and these results worried us.

The flowrate is just early volumentric flowrate that used to justify and calibrate the GC .

Currently, we are using 60 mL/min for nitrogen and 60 mL/min for oxygen.

In the near future, we might use higher or lower flowrate too in order to study the effect of feed flow rate on oxygen permeation rate (We are working on the the development of MIEC membrane for oxygen separation from air).

We are using packed column and fixed loop sampling valve.

Our GC uses argon gas as carrier gas .

The injection of the GC is through online automated method.

We hope to getting your opinions or responses.

Thank you.

Salwa Meredith Hashim

[AICMM]

mbeckwick,

Most nitrogen has argon in it as a function of the production process (1% argon in air.) As noted, sieve does not do this separation easily. So it is highly likely that your plant nitrogen has argon which makes the O2 measurement difficult. This is the main reason most ASU's use a O2 instrument.

My other questio/comment: What detector are you trying to use to measure those levels?

Sorry this does not address the GC problem at all but you are in good hands there.

Best regards,

AICMM