Oxygen Peak missing in MS

[rodrigue]

Dear all,

I have a problem trying to analise a mix of Oxygen, Methane and Argon.
I have a Agilent GC 6890N with a MS 5975B, with a Gas-Pro column.
The problem is that if I put only Methan-Argon mixture (1% concentration) is OK, I can detect the Methane peak (forget about the Argon, is only my diluent gas)

subir imagen

If I put only Oxygen-Argon mixture (6% concentration) is Ok, I can see the oxygen peak at 32 z/m and a little at 16 because of its ionization.

subir fotos gratis


BUT! if I mix both of them, I cannot see any oxygen peak. I can see the methane peak but the oxygen has disappeared.

hosting imagenes

I have already checked if the oxygen is below the detection limits, and I diluted with more Argon, (xO2=0.1%) and I can still detected.
The sample is at 25C and 1bar (absolute) so no reaction is possible.

Do you have any clue what is happening?

Thank you very much.

[rodrigue]

Has someone experienced this before? Has any idea where I can start looking for any failure?

[CE Instruments]

The 5975B is not the best for gas analysis, and you have barely given us any info regarding injection etc. The traces all seem to show peaks at 3.5 minutes do you get separation ? Why can you ignore Argon ? From the scales should you not have a huge peak for mass 40? Could the Argon be supressing the response for Oxygen ?

[rodrigue]

The 5975B is not the best for gas analysis, and you have barely given us any info regarding injection etc. The traces all seem to show peaks at 3.5 minutes do you get separation ? Why can you ignore Argon ? From the scales should you not have a huge peak for mass 40? Could the Argon be supressing the response for Oxygen ?

Thank you very much for your answer.
Sorry if I didn't give enough detail, the sample is taken on-line, we have a probe connected to a transfer line and a vacuum pump. With the "gas sampling valve" (6-way rotary valve in the GC) we fill the 1 ml loop, and then injected in the injector with the Helium flow. The injector is working in split mode at 40:1 ratio. (I have tried splitless and different ratios and still the same problem).
What I sent before was the z/m 32 and 16 only but not the complete chromatogram. If I analyse the z/m 40 as you say I found a huge peak at about 3.5 min (is eluting at the same time as O2). I cannot separate with the instrument I have now (I would need a cryogenic stuff) the O2 from the Ar, so there is only 1 peak for this two. After that I use the convolution to "separate" it with the MS signal.
I can ignore Argon because I used as a diluent, I am not interested in the Ar behaviour, composition... It is an inert gas, and it should be constant during my process.
I don't think the Ar is suppressing the O2 response because when I analysed O2+Ar only, I cannot see any problem, and I have also diluted O2 with much more Ar and I can still detect it. It is only when I mix it with CH4 when I have the problem that it disappear.

Thanks for the advice that the 5975B is not the best for gas analysis, but it is what I found when I started this job, and right now we don't have a budget for this, so I will try to live with this, but I'll keep in mind your advice in case we can change it.

Once more thank you very much for advice, as I said I am totally lost with this.

[James_Ball]

The 5975B is not the best for gas analysis, and you have barely given us any info regarding injection etc. The traces all seem to show peaks at 3.5 minutes do you get separation ? Why can you ignore Argon ? From the scales should you not have a huge peak for mass 40? Could the Argon be supressing the response for Oxygen ?

Thank you very much for your answer.
Sorry if I didn't give enough detail, the sample is taken on-line, we have a probe connected to a transfer line and a vacuum pump. With the "gas sampling valve" (6-way rotary valve in the GC) we fill the 1 ml loop, and then injected in the injector with the Helium flow. The injector is working in split mode at 40:1 ratio. (I have tried splitless and different ratios and still the same problem).
What I sent before was the z/m 32 and 16 only but not the complete chromatogram. If I analyse the z/m 40 as you say I found a huge peak at about 3.5 min (is eluting at the same time as O2). I cannot separate with the instrument I have now (I would need a cryogenic stuff) the O2 from the Ar, so there is only 1 peak for this two. After that I use the convolution to "separate" it with the MS signal.
I can ignore Argon because I used as a diluent, I am not interested in the Ar behaviour, composition... It is an inert gas, and it should be constant during my process.
I don't think the Ar is suppressing the O2 response because when I analysed O2+Ar only, I cannot see any problem, and I have also diluted O2 with much more Ar and I can still detect it. It is only when I mix it with CH4 when I have the problem that it disappear.

Thanks for the advice that the 5975B is not the best for gas analysis, but it is what I found when I started this job, and right now we don't have a budget for this, so I will try to live with this, but I'll keep in mind your advice in case we can change it.

Once more thank you very much for advice, as I said I am totally lost with this.

What is the temperature of your inlet?

Is is possible that the Methane is reacting with the Oxygen in the injection port to generate CO2 and H2O?

If you look at m/z 44 and 18 do those give peaks when you inject the mixture of O2 and Methane?

If the response for Methane in the mixed standard is lower than when you inject Methane/Ar only, it could be a sign that the reaction is happening in the injection port.

[rodrigue]

What is the temperature of your inlet?

Is is possible that the Methane is reacting with the Oxygen in the injection port to generate CO2 and H2O?

If you look at m/z 44 and 18 do those give peaks when you inject the mixture of O2 and Methane?

If the response for Methane in the mixed standard is lower than when you inject Methane/Ar only, it could be a sign that the reaction is happening in the injection port.

My inlet is at 200C.
The main aim when this is fix (hopefully) is to analyse in a reactor the behaviour of the methane decomposition at different conditions. And its needed a very long retention time (even days) to react the CH4 and O2 at this conditions (200C, 1 Bar and very low concentration). To support this, the mixture I've added is also stoichiometric, so if I run out of O2 the methane should have gone too. And the transfer line is prepare to freeze the reaction (in case is happening).
But just in case I checked my Chromatogram and there is nothing in the z/m 44 or 18. There is just a little noise where the O2 should be, but I assumed is a little amount of other compounds that the Argon has (it is Argon N50).



Thanks a lot for your comments, I really appreciate any help!

[Peter Apps]

How are you preparing the mixture of argon, oxygen and methane ?

Peter

[rodrigue]

How are you preparing the mixture of argon, oxygen and methane ?

Peter

I have two cylinders:
- O2+Ar (6% O2)
-CH4+Ar (1%CH4)

With a couple of mass flow controller I put a flow of 1 l/min (CH4 cylinder) and 0.35 l/min (O2 cylinder) into a reactor (plug flow reactor).
Right now and because of this problem the reactor is at atmospheric condition (22C and 1 bar (absolute P). I left the sample enough time inside the reactor until it gets homogenized.

Thanks

[CE Instruments]

Your GasPro column gives you virtually no separation between Methane, Argon and Oxygen and next to no retention at all, you might be better off with a Molsieve based column to at least get some different retention times. This does not however explain the issue. Looking at the traces it appears to me that your Methane in Argon gives a big dip for Argon in the 32 mass range. No such dip occurs with the Oxygen in Argon. When you mix these you get the big dip , hence my question regarding suppression of the Oxygen peak. I can't explain why one gives a dip and the other doesn't when both are Argon but it might point to a problem ?

[rodrigue]

Your GasPro column gives you virtually no separation between Methane, Argon and Oxygen and next to no retention at all, you might be better off with a Molsieve based column to at least get some different retention times. This does not however explain the issue. Looking at the traces it appears to me that your Methane in Argon gives a big dip for Argon in the 32 mass range. No such dip occurs with the Oxygen in Argon. When you mix these you get the big dip , hence my question regarding suppression of the Oxygen peak. I can't explain why one gives a dip and the other doesn't when both are Argon but it might point to a problem ?

Yes I have tried a Molsieve, in fact is already installed in the GC connected to the other inlet and to a TCD but not longer in use, I have too much Ar, so if I want to separate O2 from Ar in a molviseve I would need a cryogenic system, and right now I cannot afford it.

Yes I have suffered from this dip since I started using this equipment.But it is present in the O2+Ar measurement, it is only the scale that doesn't make easy to see it. I cannot explain why this is happening, I used to relate it with the "gas saver" function that the GC has, but I have checked that there isn't any relationship between them.
Your can see in the MS signal (without deconvolution) that this "valley" is already there. I have never found an explanation for that, but it has never affect my peaks so I have never become worried about that. But maybe is the origin of all this.


Of course your comment are always welcome, I wish we can find a solution, I have tried many things but I have a short background with GC-MS, so at this point I am totally desperate to find a solution.

Thank you very much indeed.

[Peter Apps]

How are you preparing the mixture of argon, oxygen and methane ?

Peter

I have two cylinders:
- O2+Ar (6% O2)
-CH4+Ar (1%CH4)

With a couple of mass flow controller I put a flow of 1 l/min (CH4 cylinder) and 0.35 l/min (O2 cylinder) into a reactor (plug flow reactor).
Right now and because of this problem the reactor is at atmospheric condition (22C and 1 bar (absolute P). I left the sample enough time inside the reactor until it gets homogenized.

Thanks

Rig up something where you can mix the flows in something that is not a reactor. No matter what the temperature and pressure I would be very suspicious of anything with "react" in its name.

Peter

[James_Ball]

That dip in the baseline could be related to an experiment I tried last week. I was trying to see if Nitrogen could be substituted for He or H2 as carrier gas in GC/MS. The chromatography we very good, but I lost nearly 90% of my sensitivity. It could be that when the Argon is eluting from the column it is suppressing the ionization in the source and it causes the loss of the analytes. That would also explain the dip in the baseline. Does the dip occur if you inject only Argon with no other gasses added?

[rodrigue]

How are you preparing the mixture of argon, oxygen and methane ?

Peter

I have two cylinders:
- O2+Ar (6% O2)
-CH4+Ar (1%CH4)

With a couple of mass flow controller I put a flow of 1 l/min (CH4 cylinder) and 0.35 l/min (O2 cylinder) into a reactor (plug flow reactor).
Right now and because of this problem the reactor is at atmospheric condition (22C and 1 bar (absolute P). I left the sample enough time inside the reactor until it gets homogenized.

Thanks

Rig up something where you can mix the flows in something that is not a reactor. No matter what the temperature and pressure I would be very suspicious of anything with "react" in its name.

Peter

I've already manage to build a rig to mix this two, I'll do some measurements and I'll give you some update.

Thanks a lot.

That dip in the baseline could be related to an experiment I tried last week. I was trying to see if Nitrogen could be substituted for He or H2 as carrier gas in GC/MS. The chromatography we very good, but I lost nearly 90% of my sensitivity. It could be that when the Argon is eluting from the column it is suppressing the ionization in the source and it causes the loss of the analytes. That would also explain the dip in the baseline. Does the dip occur if you inject only Argon with no other gasses added?

I cannot test this right now, I have disconnected all the possible connections to the reactor, but as soon as I finish the other test I will try and give you an answer.

Thanks a lot!

[vrodriguez]

Hello rodrigue,
It seems you may have the wrong column for this application. Please contact Agilent at: 1-800-227-9770, select option #3, then option #3 again, and lastly option #1. Or send an email to GC-Column-Support@Agilent.com so that you can receive assistance for this issue.

[rodrigue]

How are you preparing the mixture of argon, oxygen and methane ?

Peter

I have two cylinders:
- O2+Ar (6% O2)
-CH4+Ar (1%CH4)

With a couple of mass flow controller I put a flow of 1 l/min (CH4 cylinder) and 0.35 l/min (O2 cylinder) into a reactor (plug flow reactor).
Right now and because of this problem the reactor is at atmospheric condition (22C and 1 bar (absolute P). I left the sample enough time inside the reactor until it gets homogenized.

Thanks

Rig up something where you can mix the flows in something that is not a reactor. No matter what the temperature and pressure I would be very suspicious of anything with "react" in its name.

Peter

Hi

I've done this test, and everything seems to be OK, at the beginning I have had some problems with the leakage, for this take me so long to answer.
But the GC-MS can measure both peaks, methane and oxygen, without any problem.

So I have to assume the problem was in the reactor, and it seems like at very low temperature (it is prepare to work up to 1200C) the mixing is not guaranty.

So everything seems to be fix now, thank you very much for your time, and sorry I was asking you to find the problem in the GC when it was outside.