LC-MS reproducibility of MS vs DAD signals

[Colinpr]

Having a bit of a nightmare with a drug metabolite standard curve reproducibility for a new assay.
Repeated injections from a single vial give a perfect DAD signals, but the MS signals can vary by 30% !?
Have repeated the exercise with a standard solution of caffeine and experienced the same problem.
Everything (including the quad) in the MS has been removed and cleaned by company technicians - and the problem remains.

Would be vey grateful if a calmer mind can point out an elephant in the room so obvious that I can't see it, or detect a tiny but deadly trick of the evil chromatography elf.

Thanks.

[Ken]

Hi, you mentioned that this happens for a new assay; how about reproducing old assays and working standards that you know previously and worked out nicely for both DAD and MS? With this, you will be able to determine whether it's hardware or the chemistry problem ie assay, LC buffers etc.

Which, may I know what HPLC conditions and MS or MS/MS conditions that you're running the new assay?

Thanks !

[Colinpr]

Hello Ken,

Thanks for your response. I'm running a single-quad LC-MS with ESI.

The caffeine trial was done to eliminate any chemical effects from my assay preparation, since it is a standard test substance used by the manufacturer and simply run in MeOH/H2O.

Another factor was that the instrument was idle for a few months before I cranked it up for this assay. Hence the thorough cleaning that was done.

The latest idea is that the problem may be related to variability in the pressure or flow of nitrogen where it really matters. Have embarked on rebuilding the supply system and installing new external gas regulators. Will let you know the results.

In the mean time, all other ideas are still welcome.

Thanks,
Colin