Drifting signal with a 5973 GC/MSD

[tima]

Hello Guys,

Need some advice. We are having a mysterious problem with our once very reliable GC/MS. After running a batch of 100 samples the signal drifts downwards (we use anthracene 178 m/z as an internal standard but it's response drifts downwards as the batch proceeds ... this is a 36 hr batch run). The first thought was the electron multiplier is on its way out.

The signal then plateau off the the last part of the batch the signal stops drifting. Our tune has the multiplier tuning at 1550 volts so everything appears fine when we examine the tune file parameters.

Any thoughts and ideas are most welcome. BTW I have run at least 5 of these instruments (Agilent technologies) during my career and never had this kind of problem with a 5971, 72 or 73.

Regards....Timothy

[marcg]

Hi Timothy

Depending on what you're running, could it be more of a front-end problem? ie active sites on the liner or front end of the column, which could explain the gradual deterioration of your anthracene signal as you progress through the batch. If you haven't checked this, you could change the liner, inject and assess. If no improvement, cut between 50cm - 100cm from the front of the column and reassess.

I would have thought / hoped an EM would die a slower death!

regards
Marc

[cleh]

May be the filament is on it's way out. I would try running on the other filament (if that's bad, install a new one) and see if the problem persists.

[aldehyde]

Yeah could be the filament, try looking at an extracted ion chromatogram for the internal standard and see if the ionization is consistent across the peak width.

You could also have RFPA drift, there is a way to monitor the quadrupole voltage from the tune and vacuum control menu. I've never seen this occur in person but the RFPA could be putting out inconsistent voltage.. if you set a mass like 100 or 800 m/z and then monitor the voltage for 5 minutes does it drift? Probably not. Could also be the log amp, again doubtful...at the end of your tune when it checks for dark current and EM noise do you see anything abnormal?

Is it happening to all the peaks or just one? I would suspect the GC inlet and syringe before anything in the MS.

You can use the file > view tunes menu in tune and vacuum control to compare your tuning parameters to historical values.

[AICMM]

Timothy,

You have run 5 of these so you probably have done inlet maintenance and you have probably done source cleaning and your EM is reasonable value. Have you replaced any source parts? Have you dipped the quads?

Best regards,

AICMM

[tima]

Good suggestions. I haven't dipped the quads ever on this particular instrument. I did an ion source clean, did inlet maintenance (new liner, septum and gold seal), and retuned before running the batch.

We are re-running the batch after retuning the instrument with the second filament. We'll see if this makes a difference. Batch is currently running and I will check with the grad student sometime today.

How does dipping the quads make a difference?

Timothy

[James_Ball]

Unless the quads are way out of tune then it should make very little difference. If you test it at 800m/z and you get very unstable current reading then they are either out of tune or something in the quad drivers is going bad. I had a bad top board once on a 5971 that did this but the symptom was that the peak widths in tune would always drift wide overnight. Such that if I had a peak width for m/z 69 of 0.5 one morning then the next morning it would be 1.5

Are you using the turbo molecular pump or the oil diffusion pump?

Drifting lower over a long run can happen if something is loading up the pump oil and you are slowly losing vacuum, though I have seen that more when running purge and trap because of the water loading. What is your injection solvent? Methylene Chloride will pump out of a diffusion pump better than an alcohol or water will, and if it is fairly high boiling it could hang around in the vacuum chamber longer.

Have you removed the split vent line and cleaned that out, I have had those become clogged before and cause strange problems.

Hope you can figure it out. Good Luck.

[tima]

OK finally something good. After performing all kinds of maintenance with the injection port (changed liner, gold seal, o-ring and checked the column installation), we narrowed it down the autosampler motor.

This is the 6890 Enhanced Parameters AutoSample injector and this is the second time the motor gave us problems. I don't claim to know exactly why it gave us problems but after changing the motor our reproducibility went from +/- 20 to 30% to +/- 3.3 %.

Any others have problems with this particular version of HP's autosampler?

Timothy