Headspace analysis

[greta]

Hello,
I have a problem,
I'm making some repeatability test on my GC system.
I'm making an headspace analysis using an autosampler.
The speed of the gas is 35cm/sec; the split is 1:250. The temperature of the column (unpolar, 30m) is 170°C.The detector temperature is 300°C and the injector temperature is 250°C.
I'm injecting 3methil penthane (iso octane). I'm using microcaps for the sample addition in the vial.
My problem is that as concern peak height I have a CV% of 0,2-0,8% (that for me is ok), while the area CV% is between 1 and 2,5% (that for me is not ok).
Can you give me an explanation of this phenomenum?
I want to have a CV<1%.
Thanks


Greta

[Bruce Hamilton]

I assume you mean 2,2,4-trimethylpentane, not 3 methyl pentane. You need to make up your standards more accurately, and microcaps will not be suitable.

The best method may be to make a larger volume of a stock solution and dispense that accurately into vials.

Using a microliter syringe will be more accurate. Ensure you overfill the syringe, push down to mark, and then wipe the outside of the needle before adding it to the vial.

[greta]

Yes, it's 2,2,4 methil penthane. I mean that there are 3 methil group.
I have read that microcaps are very precise way to transfer liquids.

[Bruce Hamilton]

I have read that microcaps are very precise way to transfer liquids.

Fine, good luck ( you'll need lots of it ).
Don't let experience get in your path.

Bruce Hamilton

[Ron]

In many cases in capillary GC the area is more precise than the height. Is your data system integrating the peaks consistently? If there is any tailing or a small co-eluting peak the baseline may not be ocnsistent, which will affect the area precision.

[chromatographer1]

If you are overloading the column your heights may stay consistent while the area counts may change as the width of the peaks widen and narrow according to the overload conditions. This means you are outside the dynamic range of linearity for the injection conditions you have selected.

Rodney George
consultant

[Ron]

What areyour sample volume and the headspace conditions? Are you just adding the iso-octane to the vials, or is there also a solvent?

[greta]

I'm injecting only iso-octane. I'm injecting 200ul. the oven temperature is 80°C and the incubation time is 7 minutes.

The peak shape is quite good, I don't have tailings. However It's not perfectly simmetric. How can I undertsand if the column is overloaded?
Thanks
Greta

[Bruce Hamilton]

I'm injecting only iso-octane. I'm injecting 200ul. the oven temperature is 80°C and the incubation time is 7 minutes.

Is that the headspace volume into the GC, or the iso-octane into the vial?.

Iso-octane density = 0.69 g/mL, MW~114. If you are injecting that much liquid into your vials, you are injecting 27 mL of vapour into your vials.

How big are the vials?. It's possible you are overloading, depending on headspace injection volume, but unlikely with that split-ratio.

Microcaps are claimed ( by the manufacturer ) to be accurate to 1%, presumably with water. Iso-octane will evaporate quickly, unlike water, and has a lower viscosity and surface tension - so will creep around the glass edges. I'd expect the iso-octane precision to be worse than for water.

I'd make up a stock of iso-octane in a nonvolatile solvent and use that, or ( less preferred ) use a syringe and inject into a sealed vial.

[greta]

The headspace vial are of 20ml. In the vial I put 10ul of iso-octane. Then I inject 200ul of the vapours.
Thanks for your advises.
I will think about using syringe instead of microcaps.
However I can't understand why the height CV% is good and the area no.
Thanks

Daniela

[Ron]

Take a look on this forum at the postings involving headspace precision, and you will see most people posting would be very happy to have 3 to 5% precision with external standard measurements, which is basically what you are doing. 1 to 2.5% area precision is good in headspace analysis, if you want better precision simply use height, even though I have my doubts that overall you will see much better precision using height instead of area.

[Bruce Hamilton]

OK, I would inject 5, 10, 20 ul iso-octane into different vials and analyse those, using the 200ul injection. If the peak height and area do not double each time, you are overloading something - whether it be the column, detector, or headspace system.

You can play tunes with the investigation by changing the injection volume, headspace parameters, and GC/Integrator parameters.

Because you presumably have nothing else in the vials, you should get fairly consistent results, but if they vary more that you are permitted, you should consider using whatever analyte/headspace system analytical check the manufacturer suggests for assessing/confirming sampler performance.

If you meet themanufacturer's performance criteria, then you can look elsewhere for causes of variation.

[Ron]

Let's just do a quick thought experiment here. The microcaps are claimed to be accurate to 1%, so you are placing between 10.1 and 9.9 uL of the sample in each vial. The volume of the sample you place in the vial can vary by 2%, and you are seeing a CV above 1%. To me this seems to be a very reasonable CV considering the sample prep.

[greta]

Thank you very much for all your advises.
Greta