capillary GC - light gas analysis problems (hydrogen)

[rekuci]

I've been trying to get an old HP 5890 GC with TCD in disuse for 2 years working again for an analysis of H2, O2, N2, CO2, CO, and CH4. H2 is the most major and important component to quantify.

The system has been a mess...leaking everywhere (couldn't keep pressure in the carrier gas line), ridiculous noise levels (100000s in intensity) that was finally fixed after weeks of frustration by changing the TCD board. Now things appear to be at least functional, but I'm having trouble analyzing calibration gases.

The system has a 30m x 0.32 mm Carboxen 1010 PLOT column. I also have a 5A fused silica column at my disposal, which I had initial success months ago separating O2/N2 from air. Typical conditions tried are oven at 25-35C, inlet 150-200C (no purge), TCD 200-230C, 1 mL/min carrier gas, 1 mL/min makeup gas, and enough reference to get a baseline signal between 10-40 (suggested by service technician). Baseline noise is now quite good and after equilibration is flat. We can use either helium or argon as carrier gas. Is it true that for detection of hydrogen, argon is best? But wouldn't argon be worse for everything else?

When I inject 5-20 uL of a calibration gas with 4% of the 6 gases in question in helium, I see only 4 peaks, two of which are not resolved. More puzzling, when I inject carbon dioxide, I only see the same 2 peaks (O2/N2?). In argon, there is a large negative double peak after about 10 minutes (O2/N2?), and then some small positive peaks at upwards of 20 minutes after ramping up the oven. No hydrogen in sight!

What is the proper technique for injecting a calibration gas from a small cylinder? I'm using a septum fitting at the end of a mini-regulator, pulling the gas into a gas-tight syringe, closing the syringe (it can be pushed open or closed), immediately injecting into the inlet, opening the syringe, pushing the gas in, and starting the run.

What could be going wrong? Any suggestions are greatly appreciated. At my wits end.

[chromatographer1]

What is your split flow? Are you getting enough sample on column?

Helium is not a good choice for hydrogen measurement due to the "W" peak produced at ranges around 7%.

Nitrogen is the best gas for hydrogen measurement, unless you are measuring other gases at the same time where argon is a good compromise as a carrier gas.

[rekuci]

It's not a split/splitless inlet - carrier flow goes straight into the inlet and there is a purge gas outlet which I had to cap off due to numerous leaks (this was the only way I could get the inlet pressurized).

I think nitrogen hadn't been used because nitrogen is being analyzed for.

I'd be happy to even see a W peak right now, but hydrogen isn't being detected with either helium or argon carrier. I'm not sure what's happening to it.

Maybe I have some problem with positioning of the column in the inlet, it's damn hard to get that right, if only you had a cutaway like they show in the pictures.

[chromatographer1]

You wrote:

"What is the proper technique for injecting a calibration gas from a small cylinder?"

A sampling valve with a small sample loop (2-10µL) can be used to do a direct injection onto a 0.32mm ID capillary @ 1cc/min flow. Dead volumes must be minimized for this to work however. A split injection is highly recommended.

A sample loop of 100 to 250µL can be used for a micropacked column if 0.75m or 1mm ID with a flow of 5 to 25cc/min.

A sample loop of 0.25 to 2cc can be used for packed columns with flows of 10 to 60 cc/min.

I can send you chromatograms with conditions if you send me your email address.

rgeorge @ sial.com

[rekuci]

We don't have a sample loop, just a glass tube lined inlet with a top septum for syringe injection.

And this is capillary GC, not packed column...I once used a gigantic Gow-Mac GC with a sample loop and molec sieves packed column and it coulda done this gas analysis in a snap, unfortunately I'm stuck with this manual syringe injection capillary crap right now.

[MikeD]

To chromatographer1

On the Carboxen 1010 PLOT column is there baseline resolution for oxygen/nitrogen at 35 C ? I looked at a Supelco website chromatogram but it's unclear - I can only see one peak but the text says different.

[chromatographer1]

Separation in the application note is about 90% assuming oxygen is the smaller component, even with a 90% nitrogen sample and a 1% oxygen peak. The same separation is found if both peaks are at 500 ppm.

The 1010 PLOT would not be the column of my choice if N2/O2 were the primary separation of concern. A mole sieve 5A PLOT would then be the best choice.


The injection methodology appears to be Rekuci's problem here. Too much dead volume and too broad a sample plug entering the capillary.

I would recommend a very small ID inlet (on column with a 0.53mm ID FSOT tube directly connected to the capillary and try to keep the injection volume low as possible.

Otherwise you really should do a split injection to get good capillary chromatography. A 5A mole sieve PLOT would be less sensitive to dead volume as the separation is much greater for O2-N2, but then you lose the CO2, and C2s peaks.

A micropacked MS 5A could also be used at these low flow rates. I suggest a minimum of 3m 0.030" ID or 0.040" ID Carboxen 1004 80/100 at 35°C

I have chromagrams. Feel free to email me if you have additional questions.

Rodney George
Senior Research and Development Scientist
Gas Separations Research
Supelco
595 North Harrison Road
Bellefonte, PA 16823

814-359-5737 voice
814-359-5459 fax
rgeorge@sial.com

[rekuci]

The primary gas of concern to detect is hydrogen, but I'm also trying to at least identify nitrogen, oxygen, carbon dioxide, carbon monoxide, and methane. It's not absolutely crucial that the CO2/CO peak be resolved, as these are expected to be quite minor components.

I've been trying to use argon as a carrier gas, but I've been completely unable to get a hydrogen peak so far, with very low carrier flows of 1 mL/min or less and injection volumes between 10-20 uL. When I inject 10 uL of air, there is a rather poor negative double peak, but I guess the sensitivity would be pretty low with argon. What do you mean about the inlet ID? The inlet has a glass liner and you're supposed to position to capillary column inlet properly, I repositioned it because it looked a little off but still no luck with hydrogen.

I was trying out a 5A molec sieves column months ago, it definitely gave better separation of air (2 minutes and 6 minutes) at slightly elevated temperature. I want to try argon carrier with this column and see if I can catch hydrogen. Will CO2 get trapped in the 5A molec sieves? I guess a high temperature ramp would help.

Thanks again for your suggestions.

[chromatographer1]

Dear Rekuci

The results you describe are quite unacceptable and don't appear valid. This means the techniques with which you are trying to conduct the analysis are inappropriate.

You should only expect to see 1000ppm of hydrogen using helium carrier using common techniques.

I would like to help you but it might be best if you contacted me personally. You are welcome to do so.

Rodney George
Senior Research and Development Scientist
Gas Separations Research
Supelco
595 North Harrison Road
Bellefonte, PA 16823

814-359-5737 voice
814-359-5459 fax
rgeorge@sial.com

[anna]

I used to run the same gas analysis that you are trying to set up where I work. I used a Supelco Mol. Sieve 5A PLOT column (30m x 0.53mm P/N 2-5463) on a 5890A GC with the following GC conditions: Injector - 220C, Initial Temp - 70C, Ramp @ 20C to 220C and hold for 2 min.
I had good separation for the peaks which eluted in the following order on the PLOT column: Oxygen (1.59 min), Nitrogen (1.74 min), Methane (1.85 min), Carbon Monoxide (2.01 min) and Carbon Dioxide (7.54 min).
Also I would take the gas standard that I used and transfer it to a Tedlar gas bag. It was easier to work with the standard in the Tedlar bag and was more reflective of how we ran samples.
There was also one problem I would encounter alot and that was the syringe needle I used would get plugged occasionally with septa while I was doing direct injections on the GC.
Hope this is of some help.
Anna