Causes of Drift

[arch0082]

I am running an LC/MSD in positive ESI mode. My standards contain 4 components (compounds) In most runs, 1 of the 4 components drifts by as much as 20-50%. Its not always the same component either but usually just one. Once in a while 2 components drift. The UV reveals no drift to speak of 1-3%, on the same injections, the MSD drifts drastically.
So my question is, what on earth could cause such drastic drift in 1 component, and yet leave the other 3 in the range of normal operation (1-5%)? If there is something in my MSD that needs attention, what would that be? Regular maintenance has been tried such as cleaning the source and verifying the needle is installed per the manual's instructions. It passes all diagnostics I can find in chemstation, and the tune works every time.
I am using a 0.1% formic acid/ACN gradient. The only thing that we are doing which is at all outside the norm is that the injections, (not the mobile phase mind you) injected at 25 uL, contain a phosphate buffer. Meaning the diluent used to make them was made with sodium phosphate. (60 mM)

Does any one have any ideas? I need ideas as to why 1 component can drift while others stay the same.

Does any one feel that any amount of salt in the MSD is just craziness?
Everyone in my group thinks the salt is too little to be a problem, but I've definitely read that salts=bad for MS.

I am stumped so any ideas would be greatly appreciated. Even poorly thought through ideas might help!

Thanks!

[Kostas Petritis]

Is your LC method isocratic? You might be experiencing ion-suppression from "something" in your sample that is eluted after your analysis finishes and interferes with the next injection.

[lmh]

Does your standard drift when you only run standards and blanks, or does it only drift following samples?

[arch0082]

Posted: Mon Aug 11, 2008 4:58 pm Post subject:

--------------------------------------------------------------------------------

"Is your LC method isocratic? You might be experiencing ion-suppression from "something" in your sample that is eluted after your analysis finishes and interferes with the next injection. "


The method is not isocratic, the % ACN varies from 10% to 90% back to 10 % during the course of the run.


Posted: Tue Aug 12, 2008 5:02 am Post subject:

--------------------------------------------------------------------------------

"Does your standard drift when you only run standards and blanks, or does it only drift following samples? "

The standards drift regardless of whether or not they are followed up by samples. Typically, precision injections (6) drift slowly up or down, then the first set of samples or linearity standards (10 injections usually) come off, and then the first check standard comes off with crazy drift. 20-50%. but again, usually just 1 component. Most of the time its the same one.
thx for responding!

[Kostas Petritis]

Which compound of the 4 is usually experiencing the drift (i.e. is it the one closest to the void volume)? Do you send your column breakthrough to the mass spectrometer of the waste? Do you experience any retention time shifts along with the intensity shifts (i.e. inadequate equilibration times)? Maybe you should try to change your diluent to something without non-volatile salts and try again?

[arch0082]

Posted: Tue Aug 12, 2008 6:13 am Post subject:

--------------------------------------------------------------------------------

Which compound of the 4 is usually experiencing the drift (i.e. is it the one closest to the void volume)? Do you send your column breakthrough to the mass spectrometer of the waste? Do you experience any retention time shifts along with the intensity shifts (i.e. inadequate equilibration times)? Maybe you should try to change your diluent to something without non-volatile salts and try again?

Un fortunately the one that drifts the most is the third to elute, the one that drifts next most often is our IS which elutes last.
We haven't had any noticeable retention time shift. As for equilibration, we have tried using the calibration standard (5-6) injections, samples (5-6) injections, and high concentration standards (5-6) injections to try to expose the ion source to the "worst case" before testing. None of those solutions has had any effect on the drift.
Its also worth noting that this worked once upon a time just fine. We've been in method development for almost 8 months now and this problem just creeped up in the last 2 weeks which is why I'm more inclined to wonder what I might clean/check on the MSD.

As for the plumbing, the column goes to the UV, the UV goes to the divert valve, and from there to the source. We do use the divert valve to send everything to waste except for during the 15 min or so we expect peaks to come out, during that time it goes to the ion source. Thanks again for your thoughts.

[Kostas Petritis]

By the way you operate (and especially as the method worked once upon a time) phosphate shouldn't be a problem. Did you change something else in the method (i.e. column wash time in high organic etc)? Have you tried a new column? Do you ever experience the drift in the beginning of your batch (if you only get the drift after several injections, you would think that something accumulates in the column (that might not be sample related) and then start leaching)...?

[arch0082]

By the way you operate (and especially as the method worked once upon a time) phosphate shouldn't be a problem. Did you change something else in the method (i.e. column wash time in high organic etc)? Have you tried a new column? Do you ever experience the drift in the beginning of your batch (if you only get the drift after several injections, you would think that something accumulates in the column (that might not be sample related) and then start leaching)...?

You are on to something because the drift only seems to occur after some number of sample/linearity standard injections. Basic LC experience tells me that my gradient must be insufficient to get something off the column such that its causing ion suppression. But does ion suppression of just one species make sense? Also what explains that sometimes area counts go up instead of down.

Good idea about the column, unfortunately we've been trying all manner of old and new columns. Nothing seems to explain why the UV doesn't drift, and 3 of 4 components don't drift.

Also are you familiar with the idea of "baking" an MSD? I read on agilent's FAQs that humidity can cause arcing on the quadroples. Our labs have gotten as high as 68% this summer and I'm wondering if I should try that. What I don't fully understand is what the consequences of arcing would be. This problem did start in July, and our other MSD in the UK is having similar problems suddenly.

[Kostas Petritis]

It is possible that only one of your compounds is affected from ion-suppression depending on it's hydrophobicity, proton affinity, etc... (or co-elution with the compound causing the ion-suppression which here do not seem to be the case) although I would expect the drift to always be one direction (i.e. downwards or upwards and not both). Also you wouldn't be able to notice any UV drift in case of MS ion suppression...

If things were so bad to experience arking in your quadrupoles, all of your compounds would have suffered. A general cleaning though never hurt... and maybe there is something to do with humidity (what does Agilent has to say about that)? Did you imply that you are experiencing the same exactly problem on another MSD?

[arch0082]

It is possible that only one of your compounds is affected from ion-suppression depending on it's hydrophobicity, proton affinity, etc... (or co-elution with the compound causing the ion-suppression which here do not seem to be the case) although I would expect the drift to always be one direction (i.e. downwards or upwards and not both). Also you wouldn't be able to notice any UV drift in case of MS ion suppression...

If things were so bad to experience arking in your quadrupoles, all of your compounds would have suffered. A general cleaning though never hurt... and maybe there is something to do with humidity (what does Agilent has to say about that)? Did you imply that you are experiencing the same exactly problem on another MSD?

The other MSD is experiencing the same problem but on that instrument the IS is drifting, not the same component. I have to assume the problem is related because the amount of drift and timing is so similar. By timing I mean that the drift doesn't really get that bad until the first check standard.

Also my four components are quite well resolved from each other.

[Kostas Petritis]

Well, from our discussion I would:

1) play around with the post-analysis column washing conditions.
2) wait until fall where the labs humidity will go down (just kidding). But seriously, if this is really a problem (and you probably will be able to tell in a few days/months) maybe your company should try to regulate humidity for the lab. Another thing that comes with higher humidity is higher microbial growth and if your aqueous mobile phases stay around for a while you might have some relevant problems...

[arch0082]

Those are some great suggestions. Thank you very much for taking the time to think it through for me. Have a great day!

[lmh]

When you say you've got the problem on 2 LC-MS systems, are there any common factors (i.e. are you using the same column on both systems, or is it a different column)?

The reason I ask is that if this method once worked, and now it doesn't, it's unlikely to have gone wrong (accidentally) simultaneously on two different columns. This would suggest that something is now changed about the samples, or something else common to both instruments.

Are your 2 LC-MS systems in the same lab? If so, do they share a nitrogen generator? Has something happened to the gas flow rates? This can easily cause the sort of effects you describe. Agilent LC-MS instruments will object if the gas flow isn't that specified, but if someone has reduced the flow to make sure the runs don't stop, because the generator is getting a bit wobbly, then you could be getting gradually damper conditions in the spray chamber. I agree arcing as deep into the machine as the quadrupoles is highly unlikely, but a damp spray chamber with inadequate drying can reduce efficiency.

If your gas flow is sitting on the "nearly adequate" point, efficiency can wander either way.

[Don_Hilton]

Two instruments wandering off together suggests a common factor. Some possibilities have been addressed above. One other to keep an eye on is the electrical power. What else is plugged in and are there surges or sags on the line. You may be able to borrow a power line monitor, perhaps even from your local service engineer.

I remember a case of poor instrument performance at about 1/2 hour after quitting time that nobody could figure out until someone noticed a janitor dragging a floor buffing machine down the hall. Sure enough, he found a nice convenient outlet - shared with the mass spec.

Good luck.