I would suggest that a 4.5 meter Carboxen 1000 60/80 2mm ID column would perform the O2, N2, CO, and CO2 separation, depending upon the sample size. (H2 and methane is also separated)
But isn't your 6850 a capillary column ONLY GC? Installing a packed column into a dedicated capillary instrument might be troublesome, but it is possible. Bypassing the pneumatics is the easier path, but will you be using a µTCD?
You might have to install an injection site outside the GC oven but that is not a problem.
One PLOT column that will perform this analysis would be Molecular Sieve 13X. The Molecular Sieve 5A will NOT separate these four analytes, as CO2 will be trapped in the pores of the 5A.
To use a Carboxen PLOT column is possible but sample capacity is limited and the O2/N2 separation could be affected.
If you use a porous polymer column (packed or PLOT) it will require subambient temperatures.
best wishes,
Rod