I'm having an issue with the determination of Ar, O2 and N2 by GC-TCD and I'm hoping for any insight into what might be going on.
First off, here's some info on the instrument/analysis method:
-Agilent 7890B GC with TCD detector
-helium carrier gas
-cryo method using liquid CO2, where Ar/O2 are separated at 10 degrees C and N2 is detected at 25 degrees C
-gas samples are collected from field sites in pre-evacuated 12mL exetainer vials and loaded directly onto the autosampler
-we calibrate using certified gas standards, which are injected into evacuated 12mL exetainer vials, just like the samples
My problem is I can't seem to get an accurate reading for O2, specifically for field samples and ambient air samples. The GC calibrates well with the certified standards (1% O2, 5% O2, 22% O2), O2 and Ar are well separated, but I'm getting over-estimates of O2 concentrations in my samples. We're working off of the premise that O2, at least in our systems, should never be higher than 20.9%, which is the percent composition of O2 in air. When I run ambient air on our GC using the method described above, I consistently get an O2 concentration of 23%, which shouldn't be possible. I appreciate that being off by 2% is usually not a huge crisis, but we're unable to report O2 concentrations that are higher than atmospheric values - it just doesn't make any sense. I've checked the peak integration, and to my eyes everything seems good. At this point I'm not sure if this boils down to a sample collection issue or an analysis issue.
If anyone is familiar with this method for permanent gas analysis, or has any troubleshooting suggestions, I would really appreciate it.