Atmospheric Methane inconsisntent - GC-FID shimadzu GC2014

[kltowery]

Hello,

I've been using a shimadzu GC-2014 to measure ppm levels of methane in the headspace of microcosms (inoculated soil in a vial with a rubber septum). I have been getting fairly clean, consistent results until the past week, where I'm suddenly getting 'fat' peaks (a/h ratio ~6-8, rather than 4 which is what I usually see). This is troubling because these data are inconsistently-inconsistent. Meaning the fat peaks show up in one or two out of three injections for any given sample, and it differs also by the day.

on the 19th, 9/16 injections had wide peaks, and all samples had an unexpectedly high ppm of CH4 (~3.5ppm)
on the 21st, 4/16 injections had wide peaks, all samples had slightly high CH4 levels, (~2.8ppm)


I am completely stumped, as any possible cause I can think of would be something systematic, and I would expect to see a 'consistently inconsistent' change from one day to the next. HELP! Please!!

Some particulars:
I inject 200ul manually with a syringe, but am as careful as possible. withdraw 250ul, express 50ul, inject the remaining 200ul steadily, hit start, count to 5, remove needle. This has served me well up until recently

I have 80/100 porapak q & n column 2m each, ID: 3.2mm
column temp: 50*C
Inj: 60*C
N2/Air carrier: 30ml/min
FID temp: 100*C
sampling rate 40msec

All gasses used are UHP.

[Don_Hilton]

Are there other peaks in the run? And, if so, do they show a similar change? Also, do retention times change?

[chromatographer1]

Could you have any carbon monoxide in your samples to give you this 'doublet' or 'fat' peak?

Could you have septum leaks causing an flow upset giving you a pulse peak upset near methane?

I suspect the CO may be the problem before the other possibility.

Good luck,

Rod

[kltowery]

There is one other peak at about 2 min, and is low, broad, and tiny... maybe .2min wide and 100uV tall, whereas the methane peak from injecting air or standards is ~600uV for atmospheric concentrations (around 2ppm) and has an RT of .6 min.

It is my understanding that carbon monoxide wouldn't show up on an FID because it contains no CH bonds, correct? This peak at 2 min has only come from biological samples, so I assume it's trace amounts of some metabolism by-product (if anyone has an idea of what it is, I'd be curious to know!). It's far enough away from the methane peak that it is no problem, except that I have to run a 2.5 min program instead of a conveniently shorter one.

I have tried changing the septum, but no luck. I find it most confusing that the 'fat peak' phenomenon occurs interspersed with 'nice' peaks from repeat injections from the same sample! I am going to try running one injection for a loooong program to see if there are any other late-arriving peaks that could just be landing on top of the methane and making it look wide? *shrug*

thanks for the input so far!!

[chromatographer1]

N2O and ethane ethylene acetylene elute around 2 min if your C1 is 36 seconds, or it could be H2S. I proposed CO because I understood you to mean the unusual 'fat' peak was adjacent/underneath to the methane peak. There are not many things that could affect it at that retention time, but CO and O2 are possibilities.

Given a high concentration even WATER will give a peak on a FID, as it changes the conductivity across the electrodes of the detector. (and don't argue with me, 'cause it happens on a PE 900, 3920, HP5890, HP6890, Varian 3400, 3500, and 3600 GCs.)

Good luck with your unknown peak.

Rod

[kltowery]

Okay, sorry for the confusion Rod and others!

The major concern is that I'm getting sporadically wide methane peaks. Similar height, same RT as standards, 0.6s. Some of these peaks are wide enough that at ppm levels, it seems to be overestimating methane concentrations by a factor of two! (which, as I'm looking for methane consumption, is not what I want to see!)

In biological samples, I'm also getting a tiny peak around 2 min that was the mystery peak mentioned before. I had thought there might be a late-evolving peak that would sometimes coincide with my newly-injected methane, and result in the 'fatness' due to overlap. Yesterday I injected one of the samples that gave funny peaks and ran a long protocol (15 min) and did not see any additional peaks that could be overlapping with later injections, so I'm still stumped!