Chromatography Forum

I getting this on a blank sample of argon doing gas analysis:



There's a very low frequency wandering during the temperature ramp (5 to 9 minutes) with a lot of column bleed at the end at high temperature (250C). This got more severe after conditioning.

Did I toast this Carboxen 1006 column and degrade it? Could it be caused by a leak in a column connection?

I noticed Agilent has a routine for "column compensation" on the FID signal. Would using that flatten my baseline out a bit? Anyone use column compensation to negate consistent artifacts on their baseline with success?

Note the spikes at 4,11, & 13 minutes are valve switching between my Carboxen 1006 & Molsieve 5 Plot columns.

> I noticed Agilent has a routine for "column compensation" on the FID signal. Would using that flatten my baseline out a bit? Anyone use column compensation to negate consistent artifacts on their baseline with success?

You mean the one with two identical columns installed ?

-------------------
Edit:

Oh no. The one I mentioned was from an old Erba Science Fractovap. One column was analytical, and the signal from the second column (with no injection) was subtracted from the signal of analytical column by electrometer. This routine compensated analytical signal for column bleed.

But isn't "subtract blank run" equivalent to column compensation implemented by Agilent ?

I think Agilent has a way to compensate for column artifacts with "column compensation", but damned if I could get it to work today. So I bit the bullet, took both my columns out, redid/tightened all the connections and put a new Carboxen 1006 coloumn in.

Hopefully, I get a better baseline when the system comes up again.

I'm convinced that "column compensation" is nothing more than subtracting "no injection blank" from analitycal signal implemented to GC for convenience.
For me it is always more convenient to subtract blank (when necessary) by ChemStation.

Have you changed your carrier gas cylinder just before your baseline worsened ?

Your peaks are so small that you have to care about 5 pA of drift on a temperature programmed run? Wow.

Best regards,

AICMM