LCMS Inconsistent Signal Ratios compared to ISS

[ChrisLL]

I have something of an issue with our LCMS system which I haven't been able to find any explanation for. I'm attempting to run amphetamine and methamphetamine in a Waters system, using a 50x2.7 column with 2.1 uM beads coated with C18, then to a Xevo-TQs tandem mass spectrometer in MSMS mode with the ESI+ spray. With the amphetamine and methamphetamine, I'm using amphetamine-d5 and methamphetamine-d5 respectively as internal standards. I have water 0.1% Formic Acid and 5 mM ammonium formate as solution A, and Acetonitrile @ 0.1% Formic acid as solution B. The solvent starts at 90:10 initial and ramps up to 0:100 over the course of about 10 minutes.

I'll have a sample of 90:10 Water:ACN spiked with my compounds and their associated ISS, and I'll run that sample multiple times. When I do, I see the signals of the amphetamine and methamphetamine generally decrease with each run. That wouldn't be such a big problem, except that their ISS signals don't decrease proportionally. The ratio of my amphetamine signal to it's ISS is generally decreasing by about 30% (even after letting my column condition at the initial settings for about an hour or so). Since this is the same sample run multiple times, that ratio should be roughly constant, but it's varying dramatically, and non-randomly. The ratio is going down with each run.

I've tried again running the Xevo in just MS mode (no fragmentation), and then the problem seems to go away, which kind of baffles me. I am completely at a loss to explain why this is happening am hoping to get some clues for additional things to look at with my system/protocol, etc.

[bcd_GCLCMSMS]

When I do, I see the signals of the amphetamine and methamphetamine generally decrease with each run. That wouldn't be such a big problem, except that their ISS signals don't decrease proportionally. The ratio of my amphetamine signal to it's ISS is generally decreasing by about 30% (even after letting my column condition at the initial settings for about an hour or so). Since this is the same sample run multiple times, that ratio should be roughly constant, but it's varying dramatically, and non-randomly. The ratio is going down with each run.

I have a couple questions to hopefully help myself and others troubleshoot:

• Are the retention times and peak shapes consistent or changing too?
• Does anything cause the signal to rebound or does it continue to degrade with each run from where it was at the end of the last run?
• How long is a run, including warm up samples?
• Have you run a blank (spiked with ISS) in the middle of the repeated sample to check for peaks?

[ChrisLL]

I have a couple questions to hopefully help myself and others troubleshoot:

Are the retention times and peak shapes consistent or changing too?
Does anything cause the signal to rebound or does it continue to degrade with each run from where it was at the end of the last run?
How long is a run, including warm up samples?
Have you run a blank (spiked with ISS) in the middle of the repeated sample to check for peaks?

-Retention times and peak shapes are consistent
-Signals do not seem to rebound. Just continue to degrade (seemingly exponentially)
-Run is approximately 11 minutes, the remaining 1.5 minutes reset the initial condition to re-initialize the column
-I repeated this running a blank in between each run. With the blanks there is no peak observable above the noise level.

[lmh]

How do the concentrations of the internal standard and the non-deuterated version compare? Is one vastly different to the other? (in which case a change in linearity would lead to a change in relative areas).

Out of curiosity, what happens to peak areas if you just inject the non-deuterated version, with no internal standard, repeatedly?

[ChrisLL]

Sorry. Didn't have time to follow up on this for a bit.

The signals of both the compounds and their ISS both decrease, but not proportionally. The compound signal decrease can be around 50% of the ISS decrease.

I took your suggestion and ran samples with just the compounds as well as just the internal samples in them, and saw the peak areas still gradually decreasing after repeated runs on both samples.