GC/MS air water check and inlet mode

Agilent 6890 GC with 5975 MS; helium gas as carrier, 30m x 0.25 mm column. I've noticed that if I run the air/water check when in splitless mode (inlet pressure = 10.6 psi, total inlet flow = 4.8 mL/min, column flow = 1.2 mL/min, constant flow), I get high values for the nitrogen and oxygen (569.1, 161.67) but if I do the same again in split mode (50:1, inlet pressure 10.6 psi, total inlet flow = 64.5 mL/min, column flow = 1.2 mL/min, constant flow) then the values are way more reasonable (9.94, 2.21).  This would seem to indicate a small leak that is perhaps masked at higher flows?

In fact, if I set the split ratio to 0.1:1 (inlet pressure = 10.6 psi, total inlet flow = 4.7 mL/min), I do observe higher values as well (167.87, 47.57).  I leaked-checked the GC before installing the column in the inlet (new septa, new liner, new gold seal) and didn't even see a decimal of a drop over well an hour.  I have sprayed both cooling spray (1,1,1,2-tetrafluoroethane) and argon on every single point in the pathway while running in splitless mode, plus I have checked the entire pathway more than once with an electronic leak detector.

What am I missing?

Parents
  • The minimum usable total flow into a split/splitless or multimode inlet is ~20 ml/min.  Super low split flows don't work properly.   At very low flows, air can migrate backwards up the split vent and septum purge lines leading you to the false idea that there is a leak.  This is especially true when the purge valve is off and the flows are redirected in splitless.  

    20 ml/min is only 0.33 ml per second....in gas flows that is hardly any!  your try at total flow of 4.7 ml/min is only .078 ml per second -- the tiniest wisp of gas.

    Do any leak checking in split mode with 20 ml/min total flow and 1.2 ml/min column flow into the MS, the optimum helium flow into the source. 

    That air being sucked in goes through your column. If the column gets heated up above, oh, 70 or 80°C or so, the oxygen can break down the crosslinked phase and permanently damage the column. 

  • Thank you; that is very useful information.  I was pretty sure the leak wasn't at the GC side so I went back yesterday afternoon, cranked up the argon pressure unto a more focused homemade nozzle and went very slowly again over every aspect.  It appears I have a small leak at the MS side plate.  I vented, cleaned off the plate with HPLC-grade methanol, sprayed compressed air all over the seal (didn't look like it had any damage) and pumped down again.  The leak seems to be a tad bit smaller but still present, so it looks like I'll be ordering some Apiezon L grease to try to lubricate the seal, and maybe a spare seal (has been in use for almost 15 years at this point).  Hopefully that will resolve the issue.

  • Apiezon...is a handy crutch, but should be unnecessary. If you must use it, remove the O ring, wipe on the tiniest bit of grease, smear it all around the entire O ring, then wipe it ALL off.  Clean the O ring groove with swabs and MeOH at that time, too.   There will be enough left to work and not cause a big mess and other symptoms.

    A 15 year old O ring should be replaced!

    The way to install that O ring.

  • Thank you for the tips!  I'll most likely replace the seal but the grease may be useful in the meantime.

  • One symptom of too much grease are air burps.  If your scan range is above 32 they may not be visible, of course.  The air gets sucked through the grease and you may see one data point with air in it.  It can be misinterpreted as noise, electronic or ionic, and the wrong troubleshooting pathway is taken at large expense.  

    Wipe the grease on the O ring. Wipe it ALL off with a lint-free cloth, the amount left in the manufacturing pores/microchannels in the O ring is sufficient. Then reinstall.

    And good luck!

Reply
  • One symptom of too much grease are air burps.  If your scan range is above 32 they may not be visible, of course.  The air gets sucked through the grease and you may see one data point with air in it.  It can be misinterpreted as noise, electronic or ionic, and the wrong troubleshooting pathway is taken at large expense.  

    Wipe the grease on the O ring. Wipe it ALL off with a lint-free cloth, the amount left in the manufacturing pores/microchannels in the O ring is sufficient. Then reinstall.

    And good luck!

Children
No Data
Was this helpful?