carryover

[ST]

Hello.

We are trying to come up with a LC-MS/MS-method for identifying opiates in urine samples. The method we've created uses a gradient wich starts from 10%meOH/ 90%buffer (10mM NH4Ac) and goes up to 90%meOH/ 10% buffer during 10 minutes, followed by 2 min 100% meOH and another 2 min of the initial gradient for equilibration.

We have a big problem with carryover, mostly with molecules that have the largest logP-values (eg methadone and fentanyl), so the problem is probably related to that.

We have tried almost everything, changing the tubing, the rotor seal, the needle, doing blank injections, washing with methanol and other solvents (like ACN or 2-propanol) in between... We also did a blank gradient, and there was no carryover, so it's obviously a problem with the injector or autosampler.

We have tried different wash/flush-solutions in the gradient, the waters-solution (MeOH-ACN-IPA-water with 1% formic acid) some people on this forum mentioned works best, the carryover is reduced to ca 1%, but then again some peaks (eg. morphine) are really bad-looking, even if the needle is washed with a basic solution before injecting.

I’d be grateful for any ideas how to solve this problem!

[Uwe Neue]

What is the pH of your NH4Ac?

[Uwe Neue]

OK, now I have a bit more time to answer and explain where my question is coming from.

The first issue id the carry-over. This can be very specific, and you need to figure out, what wash will eliminate it. It appears that the Waters generic wash works reasonably well for you, so it is something to pursue.

After you do this wash, you have peak shape problems with the strong base morphine. This is not related to the wash, but is probably related to the fact that you do not have control over the pH of your mobile phase. For morphine to give you a good and reproducible peak, you need a true buffer. Use NH4Ac with acetic acid adjusted to pH 4.75, and your peak shape problems with the bases should go away, at least if you have a decent HPLC column.

[ST]

Perhaps the sort of buffer we use have some impact as well, but I still think the wash might somehow affect the peak shapes of morphine, because when we switch the wash solution back to methanol the peak looks good again. (we tried this a few times). But then again the carryover (not from morphine, but methadone etc) increases to around 5 %, which is quite a lot.

The pH of the buffer is 6,7, and we’re using a XDB-C8 Agilent column. While developing the method we tried different pH:s (adjusting to around 4 with acetic acid and to around 8,9 with ammonium hydroxide), but the peak shapes were best when using the neutral buffer (as we have more than 10 opiates in the analysis we had to take them all into consideration, even if the morphine peak looked better in the lowest pH).

So the biggest problem is how to simultaneously reduce the carryover and keep the good peak shapes…I guess we need to continue developing our wash solution.

[Uwe Neue]

What is the sample dissolved in?
Since you are not using a true buffer in your mobile phase, the peak shape of an analyte, especially an early-eluting analyte, can depend on the sample solvent and the sample pH etc.
Related to this: it could very well be that you have some acid left in your injector from the wash that makes you successful. If the remaining acid in the injector is the problem, a second wash without the acid would eliminate that problem.