Ireparable air leak in GC

[Rmahat]

I have been using Thermo Trace- DSQ II GC/MS system since one year. It does not pass an automatic leak check from the front panel in GC. However, vacuum in MS is absolutely fine. I checked plumbing and injector part and that are okay, I changed ferrules many times but it did not help either. Most probably the small leak is in transfer line. I am in a kind of dillemma as the vacuum in MS is fine and there is no any message in diagnostic. What is the solution of my problem? Share please !

[CE Instruments]

Firstly , when did you last run the column evaluation ? Does the K factor correspond to the column you have in the system ? Another comon reason for failure to pass the column leak check when your MS values show no leak (water and Nitrogen) is a leak in the line from an SSL/PTV injector spilt line. Can also be the in line filter has come loose.

[Rmahat]

I could not do column evaluation for this new column. The inlet looks okay as I checked both the septum purge and split valve with a bubble meter. Even the inlet looks like leak proof. I plugged off column, put an aluminium plug in the inlet nut (M4 nut), closed both valves, changed to constant pressure mode and raised pressure to 350 kPa, let it stabilize and finally set pressure to off. But there was no pressure change. Even the column is okay.

[Bigbear]

Sounds like your GC is fine. Fised silica column can crack yet not break. Try a different column.

[Yama001]

I have an OI 4660 instrument that passes any manual leak check I do, but always fails its automatic leak check. The issue in this case is a faulty pressure transducer that is not worth replacing. So in my case, the instrument is fine, the diagnostic is broken.

Perhaps kind of thing is what you are dealing with.

[Rmahat]

Thank you very much for you inputs. My problem is fixed now. Actually, somehow the firmaware version in my GC was updated so it was showing some virtual air leak. I downgraded firmaware 5.00 to 4.14 (original) and the instrument started wroking fine.

[lusi]

Hi Rmahat...I have a similar instrument and the air-water spectrum is doing some strange things.

The column test on the equipment passed and so did the MS tune leak test.

The instrument keeps all day and then over night the MZ28 rises.

I also found that the 18 and 28 were high and then after a few seconds - maybe a minute- the 28 starts to drop.

Any ideas?

[Rmahat]

Hi Rmahat...I have a similar instrument and the air-water spectrum is doing some strange things.

The column test on the equipment passed and so did the MS tune leak test.

The instrument keeps all day and then over night the MZ28 rises.

I also found that the 18 and 28 were high and then after a few seconds - maybe a minute- the 28 starts to drop.

Any ideas?

Hi! looks like your system has some leak. When you have leak in the instrument you get high m/z 28 corresponding to N2 at the same time m/z 18 corresponding to water goes down. When you keep your instrument at a temperature (say 250C) the joints can expand after sometime and let air enter the system.

I believe it's a good and effective idea to give an extra tightness to the all the joints (esp inlet and detector sides ferrule and union nut in transfer line) after a number of thermal cycle. What I do is, when I change a column or ferrules, I keep oven, inlet, and transfer line temperature at 300C and let it stand for some time (say 30 min) to allow all the metallic and graphite parts to expand by thermal conductivity. After 3o min OFF the temperature and when all the heated parts come to room temperature give extra tightness with suitable wrench. This prevents from the leak. I hope this will help you. I am sorry for my late response though.

[lusi]

Thanks for your reply.
The problem was fixed but I am not 100% sure why.
With the help of a service technician we added more calibration fluid, tightened a few places in the DSQ and cleaned the ion source.

[tlahren]

Rmahat,

I have the same GC but a different MS. Make sure you are using a new, un-punctured septum. If the septum has a hole in it at all it may not pass the injection port leak check.

Also, if you have a PTV inlet, the k-factor is incorrect after column evaluation. It is lower than what would be expected. I asked a Thermo Service tech about it and he said it is because of the smaller volume in the PTV inlet. It was not properly programmed in their column eval procedure on the GC. It is assuming the SSL and PTV have the same internal volume. I'm not exactly sure how this would actually affect the K calculation but that's the reason I was told. I would imagine this could also cause a problem with the leak checking as well.

Because the K-factor is off on the PTV I find that the programmed flow rate is also incorrect (i.e., I have no idea what my actual column flow rate is). My RTs are completely different when switching from one inlet to the other at the same flow rate.

[lusi]

Thanks will have to remember that for future reference.

[jdezeeuw]

Just make sure that when you talk about K-factor, you need to measure that isothermally. I did not get that clear in this discussion. K-factors cannot be calculated in a programmed run. I see sometimes also Plate numbers calculated in programmed runs. That's also a big "no".

jaap de zeeuw, Restek corporation

[tlahren]

Just make sure that when you talk about K-factor, you need to measure that isothermally. I did not get that clear in this discussion. K-factors cannot be calculated in a programmed run. I see sometimes also Plate numbers calculated in programmed runs. That's also a big "no".

jaap de zeeuw, Restek corporation

What I meant by "program" was that the Thermo Trace GC itself is programmed to calculate a K-factor through a feature called "Column Evaluation." The oven temp is isothermal during this evaluation. We're not actually determining it manually. This has to be performed on the Thermo GCs when changing columns etc in order for the flow rate to be accurate. Unfortunately there is no way to manually adjust it in the event it is wrong as in the PTV inlet I was talking about.

However, I hear a rumor that the newer Trace GCs allow you to bypass the Column Eval and enter in parameters manually as you would on an Agilent 6890. Not sure if this is true though.