GC-MSD: Baseline drop after higher MW peaks

[oldspecguy]

I have been having a fluctuating baseline (sensitivity) problem at higher GC oven temperatures with an Agilent 5975 MSD attached to a 6890 GC (58mx0.32 mmID DB5-MS) for over a week now. Unlike most temperature related problems, one manifestatiion (in addition to lower sensitivity at GC temps >=320C) is that the baseline/signal level actually drops below the average immediately after a GC peak elutes. It almost seems like the MSD becomes less sensitive after a peak/compound has entered the source. Chromatograms attached.

Initially, I noticed that I was having jumping/erratic baselines with one of the filaments at the top GC oven temperatures. I suspected that it may be warping and getting ready to "give up the ghost". I switched to filament 2 and retuned. Things seemed to get better, but then the same problem started appearing with filament 2, but not as often.

I thought it may be the transfer line ferrule leaking at the top GC temp, but changing that did not seem to help. I checked for leaks and have found nothing. Since it seems that it is dependent on the molecular weight (non-volatility) of the analytes - higher the alkane, the more the dip in baseline, I suspected a cold spot. All of the MS and GC temperatures seem to be operating alright, even though I have been suspecting that maybe the source or quad temperature is lower than it thinks. I am currently running it at 230C and the quad analyzer at 150 C.

I am probably going to just clean the source and hope that helps, but I thought that maybe someone else has seen this and come up with the solution.

The image here is the TIC trace of a mixture of compounds, most of them n-alkanes. The top GC temperature is 320C and the last three peaks are n-C34, n-C36, and n-C37 alkanes.

Any experience with such a problem?
Sorry about the length, but I am trying to provide sufficient info.

[/img]

[JI2002]

Do you see the baseline drop if you run instrument blank without injection? It seems that it's an injection problem (inlet discrimination)rather than MS sensitivity problem. This most likely is the case when the injector and oven initial temperature isn't high enough and it only occurs for high boilers (in this case n34, n-36 and n37).

[oldspecguy]

If I inject plain toluene, there is no baseline drop anywhere. It almost seems like the higher molecular weight compounds are somehow affecting the source volume or the sensitivity of the multiplier. Could they be condensing in the source thereby changing its characteristics?

[JI2002]

It's hard to believe those compounds will condense at 230C under vaccum.

[oldspecguy]

Well, I'm continuing to have this problem.
Yesterday, I replaced the inlet liner, 0-ring, gold seal and washer, removed 20 cm of column from injector end; checked for leaks; and reran my std. mixture including n-alkanes up to n-C36 and various alcohols, acids, aromatics. I'm still having the problem with the drop in baseline after higher MW compounds and overall less sensitivity for those compounds. Chromatographically, the column is working great - otherwise perfect peak shape even for aromatics of MW 252. The only thing I think that it could still be is:

1) The source is dirty in such a way that causes mass discrimination and a momentary additional drop in sensitivity after a peak enters the source volume.

2) The column is going bad. But that is hard to believe since the chromatography otherwise seems to be excellent.

Should I repost this to the GC side of the forum to see if anyone there has ideas?

Thanks.

[Peter Apps]

The loss of area of the heavy peaks could be due to inlet discrimination - what is your inlet temperature, and are you running split or splitless (or PTV) ?

The dips in the baseline are a puzzle. Are you running the MS with the same m/z range throughout the analysis ? What is the gas flow rate - running it too fast on a 0.32 mm column might tax the MS pumps, but why that would give a base line dip ?? - possibly ions and neutrals not getting out of the source fast enough if the pressure is too high ?.

Peter

[oldspecguy]

Thanks Peter.

Here are my conditions (Agilent 6890 GC w/ inert MSD):
Inlet: 300C (also tried 280C, no difference)
splitless mode
deactivated liner

Column: 1.6mL/min flow, Constant Flow Mode
58.5m DB5ms

Aux (transfer line) T = 300C
Source T: 230C
Analyzer/quads t = 150
EMV=1879
No excess air in source

Vaccum: 3.5 to 4.5 x10-5 (approximate range)

I also have been thinking that it is somehow related to the pressure in the source, but it is strange that it would be compound MW dependent. The autotune is reasonable, so its not like I am getting m/z discrimination. But, perhaps, since the entry point of the analytes is different than for the cal gas, there is an effect in the source volume that only affects the analytes. Hmm, I need to think about how the gas stream enters the source relative to the cal gas. The funny thing is that everything was working great, back in late September, good sensitivity, good chromatography. Then all of a sudden, I was getting jumping/erratic baseline and loss of response for high MW compounds at the top temperatures of the column. I flipped to the other filament (filament 2) and retuned and the sensitivity came immediately back, but slight dips after the peaks remained. When I switched back to the other filament 1 autotune, I would once again get really bad erratic high signals/sensitivity at high column temperatures. It really sounded like a warped filament about to go.

I have checked for leaks everywhere. It all seems ok. I am thinking of cleaning the source today, but I will first change the rough pump oil - I know it is a long shot.

One last thing, I also note that although the autotune is quite good with good sensitivity at EMV 1879, the repeller voltage is pretty much maxed out at 34.7. I tried ramping the repeller voltage and it never gives a maximal signal value (modal hump m/z 219) for the cal gas at anything less. I know that this is one indication of a dirty source, but I think that the repeller voltage on our system has always been problematic.

Additional input greatly appreciated.
Orest

[oldspecguy]

Oh, by the way, the analytes are ~15ng per component.

[AICMM]

oldspecguy,

Too much flow and too large a column id would be the first things I would suggest. The 70, 71, 72, and 73 all have flow constraints under 1 mL per minute and 0.25 max i.d. which is one of the reasons that BNA's were quite challenging. Unless you have the high volume turbo, I expect the same is true with the 75. Your problem is further compounded by constant flow mode which drives the pressure way, way up at the high temperatures. I have to say, though, that your MS pressure is not that bad but I stick to my guns on what I think is the problem. To check and see, drop the flow by half and increase the run time accordingly.

I would also note that you have a lot of baseline noise for a 15 ng splitless injection. Sure there are not any other ions in the spectra with the PFTBA valve turned off (not just air and water check)? What is your low mass cut off?

Best regards

[Ron]

I would clean the ion source. I have seen some strange things happen with a dirty repeller. You have already done a lot of work, I don't see anything to lose by doing a thorough cleaning. I would replace the filaments as well if you think they are marginal. I kind of suspect a filament issue here, but just a hunch, no hard evidence here.