Major leak - no obvious source

[mschem]

I'm troubleshooting a major leak (I assume) on our Agilent GC/MS system (O2, N2 and H2O all >1700%).

The leak seemingly appeared overnight (not after any maintenance). Running samples fine one day, next day ran a sample and couldn't see anything but very high background. Ran an air check and saw that there was a major leak.

I have used our leak detector to check all fittings and leak-checked all the lines from the gas cylinder, through the filters, to the inlet, and everything inside the oven. No trace of a leak. I'm a bit stumped, particularly since this just seemed to happen out of nowhere. Cylinders are all full, filters relatively new, no recent maintenance.

Does anyone have any thoughts about where such a large leak could have suddenly come from and any tips for finding the source of this? Thanks!

[wss]

I'm troubleshooting a major leak (I assume) on our Agilent GC/MS system (O2, N2 and H2O all >1700%).

The leak seemingly appeared overnight (not after any maintenance). Running samples fine one day, next day ran a sample and couldn't see anything but very high background. Ran an air check and saw that there was a major leak.

I have used our leak detector to check all fittings and leak-checked all the lines from the gas cylinder, through the filters, to the inlet, and everything inside the oven. No trace of a leak. I'm a bit stumped, particularly since this just seemed to happen out of nowhere. Cylinders are all full, filters relatively new, no recent maintenance.

Does anyone have any thoughts about where such a large leak could have suddenly come from and any tips for finding the source of this? Thanks!

I would try to isolate the leak in the first instance by disconnecting the column from the the MS interface (probably have to vent first) and replacing it with a blank nut, then running an air/water check (after pumping back down). This will tell you if the leak is originating in the GC or the MSD. If it's somewhere in the MSD, that might explain why you can't see any leaks through the rest of the system with the leak detector.

[James_Ball]

Another thing to check is if you have run out of PFTBA in the calibration vial. If that has become low, then the 69m/z will be very low abundance and that is what the air and water are calculated against.

One place a leak can develop that is difficult to find is in the interface tube itself. The tube is actually two pieces, with the end for the column nut being welded into a larger tube that protrudes into the MS. The small diameter tube that the column exits from inside the source is the same piece that the nut screws onto on the outside, but half way between is where that is welded into the larger tube. I have had one before leak at that weld.

If you cap the inlet end, and pull the MS away from the GC, and use some DustOff sprayed around the insulation for the interface you will eventually see a spike in the scan from the gas if there is a leak there.

[DReggio]

Most places a leak wont show up as air in the mass spec, you will be loosing your carrier gas as the lines are pressurized.

A leak at the mass spec will be sucking air in.

To check the mass spec get a can of compressed air and look up the mass spectrum for the refrigerant used.

Start a manual tune scan for masses higher than air and close the PFTBA valve. Spray the compressed air around around the mas spec mating surfaces, if you see a the spectrum for the refrigerant jump up your leaks is where you were spraying a second ago.

[James_Ball]

Most places a leak wont show up as air in the mass spec, you will be loosing your carrier gas as the lines are pressurized.

A leak at the mass spec will be sucking air in.

To check the mass spec get a can of compressed air and look up the mass spectrum for the refrigerant used.

Start a manual tune scan for masses higher than air and close the PFTBA valve. Spray the compressed air around around the mas spec mating surfaces, if you see a the spectrum for the refrigerant jump up your leaks is where you were spraying a second ago.

It usually will not show up as a massive leak, but a leak anywhere in the gas supply will show up. I always questioned it when I first began working with GC/MS, but turns out the Venturi Effect from the flowing gas will also draw in some of the outside air into the gas line and will show up in the system when the leak is in the pressurized part of the system.

[DR]

Transfer line leak?

[mschem]

Thanks everyone! I managed to find the source of the leak following some of your pieces of advice. Turns out the capillary inside the transfer line to the MS had cracked, big enough to cause a major leak but in a location my leak detector wasn't hitting. Glad it was an easy fix.