Agilent 5977 GC-MS: High O2, low N2?

[jwgcms]

Morning all!

We have a fairly new Agilent 5977 MSD and 8890 GC system, running on hydrogen from a generator. This is the first GC-MS I've worked with (and we have no other analytical chemists here) so please bear with me if I've made any rookie mistakes.

I have noticed for a while now that the oxygen level in the spectrometer seems relatively high (~13% relative to m/z 69). This is more than double the nitrogen level (typically ~6%) so I presume it is not due to an air leak. This last week I've been pumping the system down after shutting it down completely for Christmas (which I understand is not ideal for a mass spec, but the safety bods had something of a fit at the idea of leaving it running on hydrogen for a week unattended) and it's equilibrated back to a similar level (tune report attached here: https://i.imgur.com/iQUfC5E.png).

So, questions:

1. Is this level of oxygen likely to cause a problem?

2. Where is it coming from? My best guess is that it must be the H2 generator?

I've had a look around online (this forum in particular has been a great resource to solve other teething problems/questions) but it seems most people have the reverse problem of high nitrogen...

[dblux_]

...
2. Where is it coming from? My best guess is that it must be the H2 generator?
...

Hi,
Cool down oven, inlet and transfer line and set carrier flow through column to zero.

Tune, then you will have the answear if high oxygen count comes from hydrogen generator.
Let us know about an outcome.

[jwgcms]

Here is a tune from immediately before turning the gas flows and (non-MSD) heated areas off: https://i.imgur.com/uaHnYmu.png

And here is one from afterwards: https://i.imgur.com/9jPMj0J.png

As you can see, contaminants (especially nitrogen) have skyrocketed since doing that, giving the 4:1 ratio I'd expect from air. I'm not sure if this indicates a small leak, or is a normal consequence of having no carrier gas flow.

Have checked all the connections with a leak detector (with H2 flowing) and not found anything, but I don't know whether those things can pick up a really small leak. Further, none of this seems to explain the high oxygen seen in normal operation (although I suppose it suggests the carrier gas isn't the issue here).

[cjm]

There are a couple things you can do to narrow down the source of the contaminants. Ideally, I would suggest pulling the column out of the transfer line and capping it. Simply turning column flow off isn't enough as the MS pulls on the column. Wait until the MS is stable and scan. Compare O2/N2 with PFTBA valve OPEN and CLOSED and see if there's any difference. No difference means the leak isn't at the calibrant valve. If it's fixed and N2/O2 levels drop, then you're introducing the contaminants through the column.

If O2/N2 levels are still high, I would use a can of duster and gently spray around where the large o-ring sits on the analyzer chamber door while scanning for proper ions. Our duster contains difluoroethane, so I scan for ions 51, 65, 47 with PFTBA CLOSED. If you see a spike for those ions, you've located your leak.

This method works really well, even for a tiny leak. In fact, you could try this before removing the column from the transfer line, you may find your problem very quickly.

[cjm]

An easy way to check if the carrier gas is to blame: Set column flow to 1.0 mL/min while scanning in manual tune. Wait a couple minutes. Note O2/N2. Now set flow to 2.0 mL/min. Wait a couple minutes. A rise in O2/N2 would confirm contaminants are coming through the column.

[James_Ball]

What type of hydrogen generator are you using?

The best for this setup would be the ones with the Palladium membrane filter to remove moisture. I have found the simpler designs with the molecular sieve let a lot of water and sometimes oxygen through in the gas. Those are good for FID carrier and fuel, but not so much for MS work.

[jwgcms]

I've carried out the quick check as cjm suggests, and the O2 immediately shoots up to 21% when you increase the carrier gas flow. So I think that nails the source of the problem.

Looked into the generator more specifically - it's a Peak Precision Hydrogen 100, which does have a Pd membrane but seems like it's more intended for operating a FID than for GC-MS use, with "only" 99.9995% purity. Evidently that fact was overlooked back when we switched from helium, and it's aggravated by the fact there's no filter on the gas line (the Agilent service engineer told us it wasn't necessary for a generator, but I think he didn't realise it was a lower-purity generator).

So it seems to me that the first step is to install an oxygen/moisture filter on the carrier gas line (we had one with the original He cylinder, but it needs replacing) and if that doesn't work out I suppose we'll have to think about replacing the generator or going back to helium.

Thanks guys, very helpful!

[James_Ball]

I've carried out the quick check as cjm suggests, and the O2 immediately shoots up to 21% when you increase the carrier gas flow. So I think that nails the source of the problem.

Looked into the generator more specifically - it's a Peak Precision Hydrogen 100, which does have a Pd membrane but seems like it's more intended for operating a FID than for GC-MS use, with "only" 99.9995% purity. Evidently that fact was overlooked back when we switched from helium, and it's aggravated by the fact there's no filter on the gas line (the Agilent service engineer told us it wasn't necessary for a generator, but I think he didn't realise it was a lower-purity generator).

So it seems to me that the first step is to install an oxygen/moisture filter on the carrier gas line (we had one with the original He cylinder, but it needs replacing) and if that doesn't work out I suppose we'll have to think about replacing the generator or going back to helium.

Thanks guys, very helpful!

That should be more than enough purity for even MS carrier, that is pretty much the purity we use for Hydrogen or Helium from a cylinder. If you see that much O2 then either the generator is not performing correctly or there is a leak somewhere in the supply lines. If you are flowing more than the generator can easily supply that can cause purity problems, but it sounds more like a leak. I do like to keep one of the small indicating oxygen traps right before the carrier enters the GC, if it turns colors then I know there is a leak in the lines somewhere.

[jwgcms]

Hmmm, OK. In that case I'll have to go through the more drastic leak-checking steps.

By way of a benchmark, what sort of H2O, N2 and O2 levels do you generally see in normal operation? I've seen a few sample tunes where these are all on the order of tenths of a percent, is that the target to aim for?

[James_Ball]

Hmmm, OK. In that case I'll have to go through the more drastic leak-checking steps.

By way of a benchmark, what sort of H2O, N2 and O2 levels do you generally see in normal operation? I've seen a few sample tunes where these are all on the order of tenths of a percent, is that the target to aim for?

I have had as low as 0.5% N2 versus mass 69 but normally call it good if I am below 2%, even at 5% it isn't a terrible leak, and I think 5% is the mark Agilent sets for being leak free.

Tuning parameters can also make a big difference, if you are tuned more for the low masses it can exaggerate the air and water presence. Those are with helium as carrier. With hydrogen I always saw a higher water (18) but if there is a little O2 then I guess it could form water with the hydrogen carrier, but I never saw any really high air peaks.

I had one 5975 that I fought for a month trying to find the leak, it turned out to be the interface itself, one of the welds was leaking where the inner tube was sealed to one of the outer tubes and I only found it when I sprayed some Dustoff into the insulation behind the column nut where the interface enters the oven. Installed a new interface and there was almost no air at all.

[aldehyde]

At minimum you should install a moisture filter, but a moisture filter and O2 filter (or some combination filter) would be best.

Without any leaks you should be able to get <2% O2, N2, and H2O after 24 hours of being under vacuum.

Exposing the pressure controllers to water isn't great and will cause them to fail prematurely (the 8890/8860 EPC design is much more resistant to failure than previous, but it still isn't good.)

Oxygen is bad because it will cause damage to the stationary phase of the column. You'll see some column damage at even 50 ppm O2 in your carrier gas, so as leak free as possible is best when it comes to oxygen (and is why I recommend a filter in addition to trying to find the source of the problem.)

[BMU_VMW]

JWGCMS:
In this case it seems to be obviously a problem with the carrier-gas. But since you mentioned to be rather new in GC-MS I'd like to add another possibility.

If you have used Methanol e.g. for cleaning or as a residual solvent in your samples (e.g. after liq-liq extraction) you can see mass 32 as well.
It will change the pattern of the air-water check and you will not find any leaks.
Setting the injector and oven at higher temp and a higher flow for a while will 'cure' this problem.