MRM method gets more sensitive as the run goes on?!

[ManaJane]

I am working on a method to quantify a small molecule on an API 4000 triple quad using MRM. I monitor two transitions for this molecule. I have decent system precision (+/- 5%) and good linearity. So far, but my method repeatability is +/- 15% of the theoretical (not good, but not bad I guess) and my accuracy is awful. The weird thing is, the accuracy readings have a trend. What I mean is, the mass spec seems to be getting more sensitive as the run goes on. Repeated injections of the same solution (same vial or different vials) show a gradual increase in the %accuracy. Starts out near 100, then goes 110, then 120, and so on. And injections of solutions from various points in the calibration curve all have accuracy >100% (some as high as 140%).

Has anyone ever seen this kind of a phenomenon before? Any suggestions?

Thank you in advance!

[aldehyde]

I've seen this on GCMS, one possibility is you have activity on your column (in my case column and liner), as you make injections the active sites react with your compound. The response for your initial injections is lower, and then as the active sites go away you get higher and higher sensitivity until it plateaus.

[ManaJane]

That is an interesting thought. Right now I am using a Phenomenex Luna 3u c18(2) column that is brand new. I am running some consecutive analyses now to see if I ever reach a plateau in sensitivity.

[aldehyde]

The experience I am referencing was with a new column and liner, and I had purged the column with He and then conditioned it for an hour and a half. I think its more likely to see with new consumables than just out of the blue (unless something were to happen that created active sites somewhere in the system).

[carl.nott]

My experience has been with GCMS as well at an enviromental lab so I don't know how much help I can be. I've seen this before (many times) due to active sites. Can you 'hammer' the instrument with a high level standard and drop the recovery back down?

[dr_Pyrex]

I have observed some differences in sensitivity between first or second injection of standard and last standard in the batch. In my opinion sensitivity can increase because of thermal stability of interface. In my opinion 3-5 injection should be enough to reach the plateau...
Using Internal standard often solve the such kinds of problems...

[Alp]

Are you using an internal standard?
Are you using any heated auxilliary gases in the source?

[Camisotro]

Can you tell us about your mobile phase, additives, and sample matrix? And anything more about the analyte itself?

Since I don't know any of the above for your experiment, this is just a random guess... but matrix effects can cause both enhancement and suppression on ESI-MS. If your mobile phase has little to no additives and your sample matrix has electrolytes, then you could actually be getting gradual signal enhancement from a buildup of trace electrolytes in the system. Of course this is a bad thing for reproducibility; what you'd really want is to include a controlled amount of (volatile) electrolyte in the mobile phase so that trace matrix components have a comparatively small effect.

If you've already got plenty of additives in the mobile phase and you've already got a good separation of the analyte peak from the void volume, then I have no specific answer but I'd be interested in hearing what solves it for you.