GC Headspace reproducability

[kirchmsm]

I am trying to develope a calibration method for a GC. Currently I am working on a headspace reproducibilty. I have tried Iso-octane (with limited success) and I have tried DMSO (also with limited success). The goal of the project is to have an RSD of NMT 1.0%. Any suggestions on solvents that are stable in a headspace run would be appreciated. I realize that this might be a long shot. I am working on a 30 meter column.

[Don_Hilton]

How are you sampling the headspace?

[chromatographer1]

There are issues in attaining this low RSD value using headspace.

For example, many vials have a variation of more than or equal to 1%.

Your pipette accuracy is 1% for Class A and more than 1% for other classes.

Other glasswares have similar variability.

Even the best splitter design can have variation of sampling that may approach 1%.

When you add all these factors up one may consider that a RSD of 2% or less is more of an achievable goal.

If you are measuring solvents by headspace and the amount being tested is 1000 ppm then 1% of that is only 10 ppm. Is a value of 980 or 970 ppm that critical?

Or if you are measuring 100 ppm, is 97 or 98 ppm a difference that could cause a problem?

Many lab workers in a high volume sample lab may find that their RSDs of sample preparation may be at least 0,5%.

Given all these factors.... what do you think?

best wishes,

Rod

[kirchmsm]

For Don, I am trying to piece meal together different methods that have been previously run on our systems, but never for a calibration. I have run 5uL samples in 20mL headspace vial. I have tried 5ml of sample in a 20mL headspace vial. I have tried using capilary tubes, and an ependorf pippette. I have tried using other solvents to stabalize (example, I was giving the advice that water would help with my Iso-Octane samples as they seemed to be loosing stability towards the end of the run). I have always tried to run five samples. Currently the method is 10uL of Iso-Octane (using an eppendorf) into a 20mL headspace vial run on a 30 meter column, agitated for 7 minutes at 85 degrees before injection. The current RSD I am getting from this test ranges from 1.5% to 10%. Around the third or fourth sample I seem to be getting carry over or something that doubles my peak area. I am running split at a 70:1.

For Chrmatographer1, A RSD of 2% is a reasonable goal, I was just curious if a 1% was also reasonable or if was reaching. Currently I am working on 1000 ppm samples, were 2% is fine, but some of the projects run on GC's are testing for lower ppm's . My goal is a calibration method that would encompas all of these tests, but indicate if there is a system problem.

Let me be clear that my department is calibration only. All analyst have access to the GC's using a variance of testing methods and solvents. My goal is to determine if the system is in good working condition or if there maybe a problem in the system. I am currently being asked to do that using a headspace reproducability and a direct inject linearity.

My higher ppm samples are not passing by more than a 5% RSD with any consistancy. I am trying to find a more stable solvent to run with, or perhaps another way of ensuring the instrument is calibrated.

Thanks for all the help.

[Yama001]

Hmm, the Eppendorfs are primarily for aqueous measurements. Perhaps isooctane is high boiling enough to behave well but I would wonder about headspace losses as well as delivery variations given how an organic solvent will wet the pipet tip.

[chromatographer1]

kirchmsm

One inherit problem with low RSD values is your using a large headspace to sample volume 4 to 1 in your example 5mL in 20mL vial.

The high values can be caused by sample preparation issues and by the Henry's constants of the analyte and solvent chosen for reproducibility calibration measurement.

I would suggest for testing, for calibration, system suitability testing,

that you prepare a solution of 1 mg/mL solution of acetone (or MEK).

Weigh approximately 50 mg into a 50 volumetric flask holding 40mL of DMAc and fill to volume quickly after the solvent addition. Mix THOROUGHLY. Then add 10 microliter portions by Hamilton syringe of this DMAc solution to vials containing 200 microliters of water ( or DMAc ) .

This will give you sample preparations of 10 micrograms of solvent in your matrix of water or DMAc.

You should have very accurate preparations which should verify the proper heating and sampling of the headspace. This 50mL stoppered flask can be retained for additional samples over the following week, or can be accurately duplicated at any time. If you keep records of the area counts of the solvent analyte and the actual weight added to the 50mL flask you will have an indication of the loss gain of sensitivity of your GC headspace system.

You should have RSD values of approximately 1% or better.

best wishes,

Rod

[Don_Hilton]

On the addition of 10 uL from a Hamilton syringe: use a 25 or 50 uL syringe - not a 10 uL syringe drawn to maximum volume.

[chromatographer1]

And use of a chaney adapter is recommended.

Rod

[kirchmsm]

We have an autosampler with headspace syringes. Whatever method I use, I will have to incorporate that autosampler into it.

[krickos]

hi

I and many of my fellow workers are not paticulary comfy with iso-octane with DMSO for reisudal solvents. To be frank never tried to develop a method on those conditions myself but R&D had delivered several of them to us and our experiance is that especially at elevated temperature, regardless of volume in vial they can be a bit tricky (all solvents but iso octane pass RSD and Rsqr requirement). So for DMSO I suspect a lower temperature in headspace vial may be suiteble so equillibrum is not pushed to hard into gasphase.

my 2 cents on this one.

[chromatographer1]

A chaney adapter is a mechanical device that allows a hand syringe to sample extremely reproducibly.

The issues in headspace are sample preparation and reproducible heating, pressurizing, and venting.

Whether you use a fixed loop or timed injection type of analyzer.

My directions were for headspace analyzer sample preparations.

best wishes,

Rod