So we are using an Antek 8060 CLND to measure concentrations of various "pure" compounds.

A while ago, we were using one compound to make our calibration curve, let's just say it was caffeine.

We could then run other compounds against the curve, and after accounting for differences in nitrogen, get an accurate measurement within plus or minus 10% discrepancy.

Now all of a sudden we are not getting equimolar responses from the detector. When looking at a compound that has twice the nitrogens as the calibration compound, we're seeing sometimes less peak response which doesn't make sense.

We found that if you change the calibration curve compound to the parent compound you are trying to test, all subsequent testing of analogues of the parent compound read accurately, but we shouldn't need to use the parent compound as a calibration compound, we should be able to use caffeine or any other nitrogen containing compound.

There doesn't seem to be a problem with our math, so I'm wondering, does anyone know any potential sources which could cause such an error?

IE - big discrepancies in peak area responses vs compounds with differing amounts of nitrogen.

Note also none of the compounds have adjacent nitrogens or azides which the manual specifically points out can lead to problems, so...I don't know where the error is coming from.

Thanks