Melamine analysis using LCMSMS (ABI 5500)

[wan]

I was wondering if anyone has experience with melamine analysis using LCMSMS, specifically the ABI 5500 system?

I had been running up to 100 injections per sequence (A few sequences every 2 weeks). Sample has been cleaned up using liq-liq extraction and SPE.

I am not sure if my samples are making the system dirty pretty fast or if there are other reasons. Our senior scientists insisted it is my melamine samples that "dirtied" the Q0 and Q1, leading to frequent cleaning of the system.

Before we had the problem of melamine polymerizing on the spray tip. I inserted blanks between each set of samples and adjusted the position of the spray tip to avoid this problem.

My LC conditions are:
(A) 10mM ammonium acetate
(B) ACN
Gradient of 95% (B) to 50% (B) from 0 to 6 min, 95% (B) from 6.1 to 10 min.
Flow rate of 0.25mL/min

I used the diverter valve to direct eluate to waste at 0 to 3 min and from 6 min onward. I had also used a guard column with my Atlantis HILIC column as well.

Any advice or comment would be appreciated. Thanks!

-----
Wan

[Camisotro]

You mentioned in another thread that your melamine analysis is in milk. That's a pretty dirty sample matrix. Is the system getting "dirty" from the melamine for sure, or might it be the milk itself?

Are you confident that your LLE and/or SPE method is doing the job of cleaning up the milk samples? Does your method involve any centrifugation, filtration etc?

I don't have experience working with milk as a matrix but a quick search pulled up this application paper for infant milk powder, which perhaps you've read:
http://www.bdal.de/uploads/media/LCMS-49_english_02.pdf

Infant milk power was extracted in water:acetonitrile (1:4) by
sonication (30 min). The resulting extract was centrifuged
at 10,000 rpm for 10 minutes, followed by the supernatant
being removed and recentrifuged at 12,000 rpm for 15
minutes. The resulting extract was diluted as required for
HPLC analysis. Spiked samples were prepared in the same
manner with melamine and cyanuric acid added prior to
extraction at the required concentration.

So I'm guessing the water in this method is to reconstitute the milk powder, and the acetonitrile is for protein precipitation (similar methods are used for plasma as well). How does this compare to your sample prep method?

[wan]

My sample clean-up involves addition of 50:50 H2O:ACN, centrifugation, further clean up with dichloromethane, centrifugation again and SPE at the final stage.

I remember this Agilent guy at a seminar telling me milk samples are one of the worst samples to work with, regardless of clean up. Yet, other labs doing analysis of melamine in milk does not seem to have the problems I faced. So I am puzzled.

[Camisotro]

It would be useful to determine if the worst of the problems are coming from melamine itself or milk, although I suppose you would need a better diagnostic than waiting to see if it mucks up the MS =)

Have you tried post-column infusion to visualize matrix effects? You could try some repeated injections of blank solvent, some injections of blank milk extracts (i.e. put "clean" milk through your entire LLE/SPE procedure), and finally back to a few of blank solvent, and see if the baseline for melamine degrades from run to run. If you see any regions where a large matrix effect elutes, you may be able to narrow down a tracer ion that you can monitor to look for trouble more easily.

[wan]

Hi ctroster,

Sorry I got side tracked by other matters and can finally come back to this issue.

I am not sure if our LCMSMS can do post-column infusion (is it a typical type of set-up for all systems?). I never did that and no one in our lab ever mentioned it.

I would look into that. Thank you so much.

[Camisotro]

Hi ctroster,

Sorry I got side tracked by other matters and can finally come back to this issue.

I am not sure if our LCMSMS can do post-column infusion (is it a typical type of set-up for all systems?). I never did that and no one in our lab ever mentioned it.

I would look into that. Thank you so much.

I got sidetracked too. Did you ever find a resolution to this issue? (I ask many many months later...)

All you really need for post-column is a spare syringe pump and a tee. It goes after the LC stack and before the MS, and you just infuse some constant (detectable!) amount of the analyte throughout the run while injecting matrix and non-matrix blanks to see how they affect that constant signal. This is something you could read up on a bit if you're interested.

[wan]

Hi ctroster,

I read up and figured out how to set up the system we have for post-column infusion, but the instrument owner is refusing to let me do that.

There is a LCMS system I might get to use, so I thought I will try it out using that instead, when I can.

That project has officially ended, so it's difficult to convince people to let me use the instruments. :/

But thanks for following up with me! I appreciate that.

[Alexandre]

In latest Waters MSs, the post-column infusion is standard, eg Waters ACQUITY TQD, even one can add post column from 2 reservoirs.