Odd periodic peaks

[dwiggins]

We have a HP/Agilent 6890 GC with a 5972A msd and a Tekmar purge and trap. It was not being used for a long time (not quite sure how long as this is a relatively new job for me) but I was wanting to get it working and have except now I get some weird periodic peaks even on a background run. I think the problem started after I injected an aqueous sample in split mode. I did also change the msd scan range to 10-400 instead of 50-500 after that run as well but I am getting some periodic peaks at 18, 28, 32, 44 as well as 94 and 168 to name a few (I am pretty sure the 94 and 168 are from diffusion pump fluid- I think someone just pulled the plug on the rough pump without venting the MSD and that is how the fluid got in the vacuum manifold).

I have tried the trick of using the canister computer dusters to try to see if there is a leak somewhere and I can't see that there is. Also, I have switched to a different column, changed the septa, and changed the inlet liner all to no avail. The peaks are separated by about 1.5 minutes. The first peak is a huge spike of N2 with O2 as well but that then settles down to much smaller periodic peaks. I am wondering if anyone can help me. I was thinking that if it was an air leak it would not be periodic. Is this correct? Since I noticed it after the aqueous injection I was thinking that it might have something to do with injecting too large a volume for the inlet liner but if this was the case would it be periodic and would it last many "background" runs?

I will try to get some ion chromatograms posted soon.

[JI2002]

If you are using P & T, why do you inject aqueous sample directly into the GC/MS? Why not purge the sample and then thermally desorb it to GC/MS? Do you make your standards in water? Don't inject large amount of water to GC.

Because your scan range is from 10 to 400, m/z 18,28,32,44 should be the ions in the baseline. I'm not sure why you observed these ions in the peaks. Same applies to m/z 94,168, if they came from diffusion pump fluid as you suggested, they only exist in the form of elevated baseline instead of peaks. Can you post some chromatograms?

When you inject water sample into the injector, the biggest peak you see should be water peak, not air peak(m/z 28 and 32).

In order to isolate the problem, you can vent the MS, disconnect the column from MS, cap the GC-MS interface with a no-hole ferrul, pump down, then run an instrument balnk(press the start button on GC). If you still see the peaks, you have a problem in MS, otherwise the problem is in GC.

[dwiggins]

Well here is a picture of some selected ion chromatographs (from a "blank" run). As you can see there are peaks for the ions (18, 28, 32, 44 etc..). I injected the aqueous sample in the first place to see if I could see more components. Since I read that you can do aqueous injections and nobody around knew better to stop me so I did the injection. The injection was about .5 uL (I don't remember exactly off the top of my head) and the inlet liner was a taper line with some glasswool (not quite sure of the exact part number off the top of my head).

[img=http://img413.imageshack.us/img413/2263/massspecvi1.th.jpg]