The Peak area of Oxigene is shrinking from run to run

[francorrene]

Hello everybody:
This is my first message in this board.
Please be gentle.
I would like to share my problems with all of you.
I am using a GC Varian 450, it has an arrange of three columns, so the first column is a poraBond Q, the second and third column are PLOT fused silica Molsieve 5A, all these columns were manufactered by VARIAN (of course), We use He as carrier gas for the first and second column and Ar for the third one.
My problem is: The peak of Oxigene is constantly decreasing run to run the peak area of Oxigene decrease. I have no idea or explanation for this behaviour, another thing, I have been looking the flows and temperature that I need to condition my columns, but I haven't found anything.
Would any of you help this poor illiterate?

Thanks

[Davide]

Could you please specify the type of detector used.

[GasMan]

Please give concentration of oxygen and the oven temperature program that you are using. Do you find that if you do not use the GC for a few days, that the levels of oxygen go up and then start to decrease again?

Gasman

[francorrene]

Hello Friends:
Davide:
We are using TCD for all the columns

GasMan:
My standard of gases is composed by H2, He, Ar, O2, N2, CH4, CO
The certificate of our standard of gases express that the content of Oxygen is 1.993 % +/- 0.2% of uncertainty.
Temperature program
Initial T =30 ºC during 6.5 minutes (to allow the elution of Ar and O2)
A step from 30 ºC to 150 ºc at a rate of 15 ºC/min
Final T = 150 ºC during 4 minutes (to allow the elution for CO)

The working pressure for the columns are as follows:
Front Column: Porabond Q 6 psi
Middle Column: PLOT Molsieve 5A 10.5 psi
Rear Column: PLOT Molsieve 5A 18 psi (to reduce the elution time of CO)


A short story about this GC
July last year, Jaime, a friend of mine (from El Salvador) was helping me to test the GC, we selected a temperature program (it is not the same temperature program I am using now), we did a lot of readings of my actual standard, the resolution between Ar and O2 was not perfect but at least it was possible to observe clearly both peaks. The interesting part is this: in that time the peak area for Ar was at least four times smaller than the peak area for O2. One week later Jaime had to come back to his country, so I started my test by myself, but I stopped my tests because the GC was not detecting O2 in the standard.

Four weeks ago, Jaime came back and together we started to test the GC again. In the first injection of my standard the Oxygen peak was missing, in that moment Jaime told me to purge the tubing of the system of all the remaining standard of gases contained in it, I did it and to my surprise the peak of O2 started to appear again but, the peak area for oxygen got a pretty obvious reduction. I never thought that the oxygen could react with the special metals of pressure regulators.
The more concerning behavior started to appear later, we observed that the peak area for Oxygen was shrinking more and more after run. Yesterday I run my standard at 500 mBar of pressure and the oxygen was missing in the second column, it appeared in the third column yet, but the peak area was fairly smaller than the peak area I got two weeks ago.

Another notice: the peak area for all the other gases has a pretty low RSD (usually lower than 2%), even at different injection pressures, except for He (in some cases RSD is 10 %).

This has been another chapter of my story.

Any advice would be heard.

Thanks

Francorrene

[GasMan]

Can you give me more information on how your GC is set up. You talk about three columns. Does the channel that gives the problem have the Porabond Q column in the flow path?

Gasman

[francorrene]

Hello GasMan:
Thanks for your intention to help me.
I am not pretty sure if I understood the question but, indeed my GC has three columns, every column has its own sample loop, every sample loop is interconnected so when I introduce my sample/standard of gases all the loops have (theoretically) the same sample/standard. I can select which column I will use to analize samples, this is particular useful when I have to analyze biogas (it has a lot of CO2 in its composition, and can produce some damage on Molsieve 5A columns) because I can tell the GC "don't operate the 2nd and 3rd columns (PLOT molsieve 5A columns), just work with Porabond Q columns (1st column).
Just to share more about my job, I am pretending to analyze geothermal gases, to do that I designed an injection system for the gas samples.

If this is not enough for you, don't hesitate to ask me again.

Thanks

Francorrene

[chromatographer1]

The problem you have may be in the sampling system and the transfer lines, but I suspect the problem was only with your molecular sieve columns, and not the poroBond Q,

am I correct?

Rod

[francorrene]

Hello Chromatographer:

You are right, the results for porabond Q column are all the time very well, the RSD rarely is higher than 1%.

I think that my problem could be my standard of gases or perhaps the tubing system that conveys the standard of gases from the cylinder to the injection point, but I still have to carry out a few tests to confirm the quality of my standard of gases. Yesterday, I evacuated a sampling bulb and took a sample of the standard of gases, so today, I will inject my standard from that bulb and compare with my previous results.

I have planned to take a sample of air, inject it at 500 mBar and compare the results with the previous results of analysis I did on air samples two days ago.

Do you have any recomendation for me?

Regards
Francorrene

[chromatographer1]

You problem is with the molecular sieve material coating on the inner walls of your column. It needs to be fully oxidized again.

The fix?

First, disconnect the poraBond Q column from the GC and remove from the oven.

Then using AIR as a carrier gas................... yes air.

Heat the oven slowly from ambient and slowly warm the oven to 200C and hold overnight.

Then cool the oven slowly to ambient.

Replace your carrier gas to helium. and reinstall the Q column.

Now you should be fine.

best wishes,

Rod

[francorrene]

Hello Chromatographer1:

Do you have any idea about the causes of deoxidation of the inner coating of these columns
Have you ever carried out the procedure that you gently share with me?
what about the purity of the air?
can I use any kind of air? or Do I have to use a special mix?

Thanks a lot my friend.

Regards
Francorrene

[chromatographer1]

Remember what a molecular sieve is, a mineral of metal salts. If they are not uniformed oxidized they can become reduced, especially in a carrier gas flow of hydrogen, but also with nitrogen. Even samples can cause the problem. Now it may be impurities in the sieve or the glue which is holding the particles to the fused silica surface of the tube that are reduced, but they react with oxygen in your sample, giving you variable results.

I have done the process I described with packed columns, but never capillary columns. You should use clean air, and you can lower the temperature to 190C if that makes you more comfortable. I rarely used capillary columns due to their inherit low sample capacity. My real world applications always used packed columns. But I prepared packings for use and know the problems pretty well.

Best wishes,

ROd