Chromatography Forum

Hi all!

I'm facing a mysterious fact.
Trying to quantify amphetamine with two different cromatographs, a Thermo 1300 and an Agilent 6890, with FID detector. Apparently, the analyte is soluble in my solvent (a 50-50% mixture of ethanol and methanol, and then 10 min of ultrasonication).
When I inject the calibration points on both chromatographs, it seems I have two different responses (same acquisition method, or that's what I think) :
for Thermo I could have an acceptable linear response (r and r2=0,99..) but the quadratic curve fits PERFECTLY.
However, for the Agilent, the linear fitting is already pretty cool, r and r2 are 0,9999...

Is that even possible? It is the same preparation.


The acquisition characteristics are the same except for the detector flows, bcs Thermo recommends another flows, but I don't think this is the source of the problem since other analytes don't show these differences.

I am pretty disappointed since I thought it had an acceptable lineality with a linear curve since I have realised that it was not perfect because its more quadratic than lineal.
I would like to avoid a quadratic fitting... Could it be due to a poor solubility of the analyte in my solvent? Do you think that changing the solvent (100% methanol could work better but I was avoiding it due to its toxicity) would improve the linear response?
(But still with the mistery that in the Agilent is lineal...)

Let's see if some of you has any idea

L

Since the same preparation gives two different types of linearity curves on the different makes of GC, I doubt that it's a sample preparation issue. However, I typically avoid using denatured ethanol for such preparation, as the denature formulation/amounts can vary if different supplier's ethanol is used. On the other hand, I DID use denatured ethanol (5% methanol, 5% IPA) as HPLC mobile phase for one previously difficult separation, but after fighting with pointy-haired boss about that product being consistent from lot to lot, we did cGMP-validate that pharmaceutical assay.

So I suspect inlet issue(s) for your discrepancy.

Hi!

Thank you for your rapid reponse .

We are using absolut ethanol, wich I think isn't meant to have any other compound in its composition (at least in these proportions you mention). Therefore, if I have understood well from your explanation, it shouldn't be a source of problem.

I don't understand (or I don't want to understand ) what you mean by inlet issues... I set the same temperature, same flows, same liner in both machines... "other issues" would imply inlet some characteristics that don't depend on me but on the chromatograph architecture? That would mean a more difficult problem to solve...
Thank you again for your words.

L

"Anything" could include...
different autosamplers
different syringes
different compositions of inlets
different geometry of inlets
different inlet liners

I see... things that are not "available" to change. Well, thank you for your time, I have been doing more tests and it seems that increasing inlet temperature is helping. Setting a new method for a new analysis takes more time than I thought

I see... things that are not "available" to change. Well, thank you for your time, I have been doing more tests and it seems that increasing inlet temperature is helping. Setting a new method for a new analysis takes more time than I thought

Sometimes a lot more time

There can be differences in inlet temperatures between instruments, even of the same manufacturer. Unless you are using a good reference thermometer and calibrating the two instruments to be exactly the same you may need to vary the temperatures between them to get a match. Also the different manufacturers will have the heating element in slightly different spots compared to the housing so one may be hotter at the top of the inlet and the other at the bottom or middle, every thing like that will cause a slight difference.

Diameter of liner, shape of liner, how fast the syringe injects, all can make a difference, so best to vary the conditions between instruments until the method is optimized.