LCQDUO quantitative problems

[dr_Pyrex]

Hello all,
I determine corticosteroides (dexamethasone, betamethasone, flumethasone, triamcinolone acetonide and prednisolone) residues in animal tissues with LC-MS (Agilent 1100-LCQDuo: ECI (-), MS/MS mode). My experience in mass spectrometry is not big so I have some questions.
I injected on my system dexamethasone standard (0,025 ug/ml) and CV (%) of peaks area was about 7-10% (six injection from the same vials).
In my opinion I checked all parameters (gases, amount of scans, maximum injection time, SIM mode, MS/MS mode, etc.).
I heard about problems with quantity with ion-traps but in my opinion CV (%) 10% is too much! (This is not autosampler problem)
Do you have any ideas?

Best regards!

[Uwe Neue]

you can check if the issue is the MS or the LC by using a UV detector at the same time. If the reproducibility of the UV response is better, the issue is definitely on the MS side. You may need to change the settings of the MS.

[Kostas Petritis]

I am not supriced with your results, but this is something specific with the ion-trap you are using (i.e. Thermo LCQDuo). Later ion-traps have good automatic gain controls that gives better CVs. I have heard from the Thermo people themselves to admit that their LCQ Duo have some issues...

[dr_Pyrex]

My question is: which of parameter of LCQDuo can be change to increase precision od analysis?

[Kostas Petritis]

I don't think there is some way other than maybe introduce internal standards or change to a quadrupole instrument or a later generation ion trap.

From the moment that you had the same MS conditions for all the 6 injections the precision should have been good/better (or as good as it gets for MS, remember that you won't achieve the precisions of a UV detector).

My point was that the ion trap you have it is more difficult to achieve good precision. I wish I remembered what the exact problem was (I think it has to do with the automatic gain control but I am not sure).

[joel.ch]

I don't think there is some way other than maybe introduce internal standards...

I would also suggest you the use of an internal standard, then repeat your experiment considering area ratios instead of areas, it should lower your CV.

Hope this helps

Joël

[dlorenzo]

I've worked with an LCQDuo and have the same reproducibility problem and then I wrote Finnigan. They said that cv with MS detector are not like UV ones, and a cv about 7% would be fine.

Hope it helps you.

regards

DL