Step in baseline, Mol sieve Column, H2O contaiminated?

[smoothy]

I have a Agilent G2890A mirco GC used for the gas analysis of Fuel Cells. I am having trouble with one of the channels. It seems that one of the columns is producing a consistent step in the base line, which looks like a step off the back of the lend peak (rapid decrease in detector signal to baseline). Since this step has appeared the column seems to produce incorrect results, despite being calibrated using standards. The column is a Mol. Sieve 5A 10m 13X . I have tried baking it out at 160°C, overnight numerous times, with 80 psi of carrier gas, which has not removed the step. 160°C is the max temperature allowed by the software, I cannot remove the column to bake in a separate oven, since the columns are in closed unit modules. Recently the column may have been exposed to a large amount of steam in the samples, however baking out in these condition has no impact.

[Rafael Chust]

Which detector are you using?

If it is TCD, excessive water coming out might condensate in the filaments, causing this problem. Increase detector temp and this might solve the problem.

[arie2044]

Hi
did you checkd for CO2 in your samples?
Mol Sieves absorb the CO2 and behaive strange after that. to prevent water effect work at 140-160 deg C at the oven and 250 at the TCD
Good Luck