Intermittent Erratic Baseline

[gb_gcms]

Hello everyone,

I'm new around here, and fairly new to GC/MS, so please call me out for any omissions or oversights. The system experiencing an issue is an Agilent 7890A GC, 5975C MSD in positive CI mode, helium, currently measuring glucose pentaacetate. Been running this protocol and method for many years.

A few months ago, I started observing small, infrequent disturbances in the baseline, sudden rises and falls that look like steps. (see image below). The issue has gotten worse, but frustratingly intermittent. I can run 100 samples with no problem then suddenly the noise will become so severe that the data is unusable, then a few samples later the problem is gone. (see image below) The baseline issue happens even when I run air and methanol blanks. I disconnected the column from the MSD, and the problem still occurs, but when I switch out the CI source for an EI, the issue seems to go away. At first I thought it was the electron multiplier, but we're tuning at ~1600 V. Another factor: our filaments have started to burn out after two weeks, and they used to last for months. Tune shows no obvious sign of a leak (autotune is fine, no peak at 32) but any advice for doing a more thorough check would be appreciated.

There is black build-up on the source body near the filament that we can't completely clean off. (see image below) Also, the repeller insulator looks dirty, but everyone is too afraid to clean it if we're not sure it's the issue (those things are not cheap..) I also noticed black buildup in the metal tube that holds the source in front of the quadrupole, near where the filament is. I'm not sure if this is a result or contribution to the issue, but it has looked this way for a while.

I've checked all the connections, system is plugged into the same outlet it always has been, no new equipment nearby. Two other GC/MSs (one CI one EI) working fine. Replaced all the usual consumables on the GC side, swapped out the CI interface tip seal.

Any idea what's going on?

[Peter Apps]

Leaking current through the dirty insulators has to be a likely suspect, but that chromatogram really does look like a scale change - if one of your peaks comes out during a baseline dip is the peak area the same as it usually is ?.

Are you monitoring SIM or full scan ?

Peter

[gb_gcms]

Leaking current through the dirty insulators has to be a likely suspect, but that chromatogram really does look like a scale change - if one of your peaks comes out during a baseline dip is the peak area the same as it usually is ?.

Are you monitoring SIM or full scan ?

Peter

If a peak comes out during a baseline dip, it tends to decrease in area. We are running in SIM mode.

This is what our insulator looks like, for reference:

[Peter Apps]

When you plugged the column interface and still saw the baseline dips was the MS running the programme that it runs during analysis (I don't use that instrument so I do not know if that is even possible) ?. If not than it rules out a whole host of possible causes and strongly suggests a problem with the signal processing electronics.

Peter

[gb_gcms]

When you plugged the column interface and still saw the baseline dips was the MS running the programme that it runs during analysis (I don't use that instrument so I do not know if that is even possible) ?. If not than it rules out a whole host of possible causes and strongly suggests a problem with the signal processing electronics.

Peter

Thanks for the help, Peter. I believe the only parameters that would impact the chromatogram with the column interface plugged are EM volts and length of time the filament and detector is on (someone correct me if I'm wrong). Either way, we tried another method with lower EM volts programmed, and the baseline disturbance was still observed. It usually takes about 10 runs until the issue starts appearing, even with the column interface plugged.

[Peter Apps]

If you are running SIM a shift in mass calibration will change both signal and baseline (although there is no particular reason why the baseline would always drop). If you have a time programme for when you look at which fragments you might see a shift-related drop only for some of the masses - but that time programme would not be running unless you are running the acquisition programme on the MS. Certainly if you are not running the programme, but the problem persists, then nothing that changes during the programme is the cause.

Is it possible you have dirt on the quadrupole, causing mass calibration to be intermittently lost ?

Peter

[gb_gcms]

If you are running SIM a shift in mass calibration will change both signal and baseline (although there is no particular reason why the baseline would always drop). If you have a time programme for when you look at which fragments you might see a shift-related drop only for some of the masses - but that time programme would not be running unless you are running the acquisition programme on the MS. Certainly if you are not running the programme, but the problem persists, then nothing that changes during the programme is the cause.

Is it possible you have dirt on the quadrupole, causing mass calibration to be intermittently lost ?

Peter

Ah, I just realized I misspoke in my original post (edited to reflect the change). When I swap the CI source for EI with the GC interface plugged, I can't seem to get the issue to occur, however, I'm hesitant to be sure because the issue is intermittent.

The SIM parameters used to test both CI and EI had no time program that I can see, I just monitor for all of the selected ions for the same amount of time. Let me know if I'm missing something here.

Dirty quadrupole is certainly a consideration, however, I'm not sure it would explain the large noise spikes and filaments burning out quickly. I don't have enough experience to feel comfortable opening up the quadrupole and trying to clean it, but I took a peek and I don't see any obvious dirt.

Does this look like something that could be originating from the source? I'm considering swapping sources with the CI that's working, however, we run a heavy operation here and that one's been running nonstop for 4 days to make up for this one being down. So I'd rather not disturb the working CI if the source can't possibly be the issue.

Thanks again for your help.

[Peter Apps]

If the problem occurs with the CI source and not the EI source, as you now say, then it must be something to do with the CI source or whatever else is different between the two setups. A really thorough scale and polish is a good place to start.

Peter

[gb_gcms]

Great news! I think I've fixed this issue. I hope this can save someone else some time/money/headache in the future.

I switched out the repeller insulator with a new one, and the issue has gone away. It seems a very dirty insulator can cause electrical disturbances that disrupt the baseline and damages filaments.

Thanks for all of your help!