Chromatography Forum

Been tracking down a leak in our 7890a GC, 5875C MS setup.

Have replaced practically everything at this point (septum, liner, o-ring, ferrules, gold seal, even the column).

Still, the leak persists.

I'm pretty confident that the source of the leak is in the inlet, but I'm unable to be 100% sure because we don't have a helium detector on hand (I'm trying to borrow one for a day or two).

When I use the Service Mode Front Inlet Leak Test (10 psi), it passes the test every time because there is no pressure drop from Inlet to Column.... but instead, I'm seeing an increase from inlet to column.

The machine is measuring total inlet flow at approximately 0.8 ml/min of helium, and the total column flow at ~0.99 ml/min of helium.

Does anyone have any ideas why this might be the case?

Have I missed something obvious?

Thanks in advance.

This is not a significant difference for the front inlet leak check. The calculated pressure of 0.99ml/min is always an approximation.

You should be suspicious if there's a bigger absolute difference, depending on the pressure. It's better to test at higher pressures like 30psi. In my setup, I know something is wrong when the absolute difference is >1ml/min at 30psi.

Is there something else that points towards a leak? Like air in the MS?

Our tunes/air and water checks have had higher Nitrogen and Oxygen values in the 4:1 ratio.

My understanding of the front inlet leak check "test" was that it was comparing the total flow into the inlet (based on pressure) and the total flow out of the column (again, based on pressure).

So, I wasn't concerned that the column flow wasn't 1. I'm concerned that I'm getting more flow out of the column than I'm putting in to the inlet.

I assume that with a significant air leak in the inlet, the total flow into the inlet might be 1.0 ml/min @ 10psi, but the column flow would be significantly less than 1.0ml/min, indicating that some carrier gas wasn't making it into the column.

My situation is reversed... I'm reading a greater column flow than inlet flow.... with a known air leak somewhere!

Do you think that a significant leak at the GC - MS interface might inflate the column flow calculation?

My understanding of the front inlet leak check "test" was that it was comparing the total flow into the inlet (based on pressure) and the total flow out of the column (again, based on pressure).

So, I wasn't concerned that the column flow wasn't 1. I'm concerned that I'm getting more flow out of the column than I'm putting in to the inlet.

That's not what i said above...

The test gives 2 values:

- One is a real-time measurement of the total flow in the inlet after the GC puts itself in a sample prep state. In short, all the flow measured here should be going trough the column. In your case 0.8 ml/min.

- Two is a calculated value, not measured, of the flow that 10psi should give based on the dimensions of the column you put into the software. In your case 0.99 ml/min. It is basically the flow you want in an ideal case trough the column.

In my post above i tried to explain that the difference between 0.80 and 0.99 is too small to be suspicious.

If you have a leak around the inlet the measurement value >>> calculated value.

Our tunes/air and water checks have had higher Nitrogen and Oxygen values in the 4:1 ratio.

I suggest the following. (Forget about the leak check test.)

Put the MS in full scan mode between 10-100 m/z. Use a dust spray remover to carefully spray at the connections close to the MS while checking if there's a signal appearing. Do the same for the door that tightens the MS or anything that you suspect sucking air.

Hello

To check inlet for leak do simple leak test:

1.Remove column and plug inlet
2.Plug Split vent line
3.Switch off septum purge
4.Change inlet to Splitless mode
5.Pressurize to 25psi and total flow 60-100 ml
6.Leave for 5 min
7.Swith off pressure (do not set 0 psi)

Check pressure drop - it should be less than 2psi in 10 min.
If it is higher you have leak.

Regards

Tomasz Kubowicz

Rndirk, thanks for your patience in explaining that. Apologies for making that complicated.

I actually played around with a can of air yesterday in a similar setup to what you described. I did not, however, test the MS door.


Worst case scenario... if I can't track down this leak, my foreline pressure is ~51 mTorr... am I hurting the machine by leaving the Quad and Source hot overnight, with ~1% H2O, ~15.5% Nitrogen, and ~4% Oxygen?


I expect the water to come down below 1% today. It was very humid yesterday when I vented the machine to work on it.

Thanks again for your help and patience.

Double check that your column configuration is correct in your method. I've seen strange things when the column dimensions are incorrect.
You can test the column flow by injecting air and monitoring M/Z 28 in manual tune. Use the stop watch function on the GC . For a 30M 0.25 column @ 1 cc/min it should come out at 1.1- 1.4 min.

Rndirk, thanks for your patience in explaining that. Apologies for making that complicated.

I actually played around with a can of air yesterday in a similar setup to what you described. I did not, however, test the MS door.


Worst case scenario... if I can't track down this leak, my foreline pressure is ~51 mTorr... am I hurting the machine by leaving the Quad and Source hot overnight, with ~1% H2O, ~15.5% Nitrogen, and ~4% Oxygen?


I expect the water to come down below 1% today. It was very humid yesterday when I vented the machine to work on it.

Thanks again for your help and patience.

The values in % are quantified towards the calibration gas. 4% oxygen is not that high, but it depends on your system. For our latest GC-MS/MS, we always manage to keep it <<1%, but for some older systems 4% is a normal value; nothing alarming. Especially if it was vented a day ago.

I wouldn't worry too much. You might be looking for a leak that isn't there. Still the check i described above (with the spray can) takes a minute so i'd just do it to be sure.

15% nitrogen versus m/z 69 is a little high, but not a large enough air leak to ruin the instrument if left hot, I have had to run samples with leaks worse than that though I don't like to do it.

On the Agilent 5973, 75,77 instruments another place you can get a leak that is difficult to find is in the metal part of the interface itself. I found mine by spraying the Dustoff into the tiny gap between the oven end of the transfer line and the ceramic insulation, it then spiked a signal from the duster gas. Replacing the O-Ring seal of the interface to the analyzer didn't solve the problem and it was confirmed by an Agilent engineer that others have had leaks occur in the interface itself, possibly from a microscopic crack at the internal weld. Mine is leaking at about 30% nitrogen. I can run a while on source cleanings but need to clean the oxidation from the repeller quite often. Just waiting on the replacement part now.