I am having an issue with my 490 microGC for analyzing H2/CO2 gas mixtures. My GC is calibrated with a 5 point calibration curve of various mixtures of H2/CO2/N2. I am using a molsieve and poraplotU column from Agilent with argon and helium carrier gases respectively. Basically, when analyzing H2/CO2/N2 gases I have no issues reading the proper composition. However if I remove the nitrogen and only analyze CO2 and H2, the compositions fluctuate by about 2-4% conc. Here is an example:

Sample 1 conc. 40% H2, 20% CO2, 40% N2. GC will the composition correctly.

Sample 2 conc. 80% H2, 20% CO2, 0% N2. GC will read 82% H2, 24% CO2, 0 % N2.

On our GC, 80% H2 is outside of our calibration range, however I still feel that CO2 should be measured correctly, regardless of what conc. of H2 is measured.

How come the removal of N2 is changing the GC's ability to measure CO2 accurately?