Permanent gas analysis by GC with TCD

[tylermarcet]

Hi everyone,

I'm trying to analyze permanent gases by GC, but have had little luck. I'm using an HP 6890 GC equipped with a TCD, and I have a split/splitless inlet with a gas sampling valve and 0.25 mL sample loop. I'm using a Supelco Carboxen-1010 PLOT column, 30 m x 0.32 mm. My problem is that, regardless of inlet parameters, I'm unable to get even a nitrogen peak when injecting ambient air samples. Using the column verification parameters that shipped with the column, I'm unable to even get peaks, much less reproduce the chromatogram. Does anyone have ideas or parameter suggestions?

I should note that this GC setup has been verified using the Agilent checkout solution and method (3 compounds in hexane gave me the 3 anticipated peaks + solvent peak).

Thanks for your help.

[aidnai]

I take it you are injecting the 3 compound sample into the split/splitless inlet by syringe? If so, my suspicions would tend to be directed at your sampling valve.

By the way, how do you get the ambient air into the sample loop?

[tylermarcet]

Yes, the checkout sample was injected via syringe.

As for loading the sample loop, there's a luer lock connector attached to the GSV, so while the valve is in the "off" or "load" position, I connect a gas tight syringe to the luer lock connector and push the sample through to fill the loop.

[tylermarcet]

Also, I forgot to mention, I've tried large (~10 uL) manual injections of ambient air using a gas tight syringe instead of the GSV/sample loop, and still haven't seen any peaks.

[GasMan]

What is your split ratio?

Gasman

[tylermarcet]

I'm running in splitless mode. Purge flow of 60 mL/min at t = 0.75 minutes.

[chromatographer1]

Do you have the proper valve configured in the GC with the on-off switches set for both activation and fill operations?

I think you may have a big leak or your valve is not moving as and when it should.

Rod

[tylermarcet]

Do you have the proper valve configured in the GC with the on-off switches set for both activation and fill operations?

I think you may have a big leak or your valve is not moving as and when it should.

Rod

The valve had come out of adjustment, though we were having the same problem (no gas peaks) when doing 10 uL direct injections before the valve was even installed. I just readjusted the valve and I'll be doing another valve injection shortly to test.

All of the leak/pressure/flow checks we've done have come out perfectly, though, so I don't believe a leak is the problem.

[chromatographer1]

Then

I think you may have a big leak or your valve is not moving as and when it should.

Rod

[aidnai]

10uL is not very big compared to your sample loop (250uL). Also, when you use the luerlock connection to put air in the sample loop, how much air are you pushing? normal 1/16" tubing has an internal volume of about 400uL per meter, so if there is a foot of tubing and you are using a 25uL syringe, you wont be getting the air to the sample loop at all (not to mention you won't come close to filling the loop).

When you do a syringe injection of air, do you get no peak at all? or just a tiny peak? And, did your checkout mix of three components show reproducible peaks, or were they highly variable?

[tylermarcet]

10uL is not very big compared to your sample loop (250uL). Also, when you use the luerlock connection to put air in the sample loop, how much air are you pushing? normal 1/16" tubing has an internal volume of about 400uL per meter, so if there is a foot of tubing and you are using a 25uL syringe, you wont be getting the air to the sample loop at all (not to mention you won't come close to filling the loop).

When you do a syringe injection of air, do you get no peak at all? or just a tiny peak? And, did your checkout mix of three components show reproducible peaks, or were they highly variable?

When running gas samples I get no peaks at all. It just looks like I'm running blanks. I'm actually using a 50 milliliter gas tight syringe, so the vast majority of the sample I push through exits the loop via the "out" line. I typically push through at least 30 mL.

The checkout mix was entirely reproducible every time.

[GasMan]

We recently had a brand new valve where the rotor was incorrectly marked so that the rotor was in fact operating backwards. This meant that instead of filling the sample loop when the valve was in the OFF position, you were actually filling the groove that connects the ports together, which is probably just a few microliters. You can easilly test if this is the case. Turn the valve ON before you start the run and fill the loop. In the run table, turn the valve OFF at the beginning of the run. This assumes that you have the valve configured as a switching valve and not as a gas sampling valve. If you now get peaks, then this is your problem.

Was this 6890 supplied with the valve installed at the factory or has this valve been added by yourself ?

Gasman

[tylermarcet]

We recently had a brand new valve where the rotor was incorrectly marked so that the rotor was in fact operating backwards. This meant that instead of filling the sample loop when the valve was in the OFF position, you were actually filling the groove that connects the ports together, which is probably just a few microliters. You can easilly test if this is the case. Turn the valve ON before you start the run and fill the loop. In the run table, turn the valve OFF at the beginning of the run. This assumes that you have the valve configured as a switching valve and not as a gas sampling valve. If you now get peaks, then this is your problem.

Was this 6890 supplied with the valve installed at the factory or has this valve been added by yourself ?

Gasman

The valve was installed by an Agilent certified engineer less than two weeks ago on an existing 6890 we had, and I've yet to see any gas peaks.

Interestingly, I followed the actuator adjustment guide provided to me by Agilent, and now the system will build/hold pressure only while OFF! This means that the system holds pressure just fine during the sample loop loading phase, but depressurizes and total flow drops to zero when the valve actuates and is in the "ON" position for column loading.

All this tells me that the valve was a) faulty from the start or b) incorrectly installed... I can't imagine it's normal for a 2 week old valve to come out of adjustment after 2 weeks of very light use! Any other thoughts?

Since the valve was installed by an Agilent certified engineer, I found out that the installation is under warranty, so another engineer will be stopping by this week to take a look.

Thanks to everyone for your replies so far!

[Johnny Rod]

Odd one, maybe the valve isn't plumbed right. I'm still surprised you aren't seeing any peaks with manual injection, but they should fix that for you as well. As said a 10ul injeciton is small but should be enough to give you peaks as you are looking for large %ages. Helium carrier gas?

[tylermarcet]

Yep, helium carrier.

I've spent countless hours on the phone with Agilent trying to figure out the total lack of gas peaks without any luck. I've been hoping for months that someone would point out some stupid mistake I'm making, but no luck.