Headspace linearity issue

[elutionorder]

I am working with a HT250D autosampler, which has a heated syringe, attached to a 6809N. I am trying to develop a PQ method for linearity.

I've been using a 5m x 0.1 mm i.d. uncoated column since this type of column has been used by engineers who've performed previous OQ's. My sample is 10 ppm ethanol in water where 2 mL of the sample is in a 10 mL headspace vial. The injection range is 0.4, 0.8, 1.2, 1.6, and 2.0mL. One of my better runs gives an R sq. value of 0.966 for the 5 points. For the first 4 points, linearity is 0.997.

Area (pA*s)
4539.04004
6119.06641
7547.75293
8765.08496
9183.38965

I'm getting close but not close enough. I've change liners, flows, splits, injections speed and haven't improved. I've increased the flow (0.7mL/min) to increase the inlet pressure to get a more compact sample plug. I decreased the injection speed to 8 mL/min keeping it to a little less than half of the total inlet flow.

Headspace: Oven 70C, Syringe 75C, Injection speed 75 mL/min.
heating time 15 minutes

GC:
Straight 4 mm i.d. liner (Agilent 210-3003-5)
Inlet temp 150C
Split 30:1
Inlet pressure 30.46 psi
Total Flow 21.4 mL/min.
Column flow 0.6 mL/min.
Oven 40C for 2min., ramp to 150C to burn off any water vapor in the column

I can provide other conditions if needed. Any input is greaty appreciated.

[JGK]

Been a while since I did any GC but I found that when working with gaseous samples it was always better to maintain a constant injection volume.

I wasnot performing headspace work and was manually injecting gaseous sample. but when varying the injection volume linearity was poor. once the volume remained consistent linearity improved considerably.

[GaryR]

As suggested by JGK I would be varying the concentration of my standard solutions and maintaining the same injection volume. This should give much better linearity results.

[elutionorder]

Thank you for the feedback. After tweaking column flow and injection speed, I was able to get good linearity above 0.99.

[elutionorder]

I spoke to soon. I couldn't repeat it. The first run was 1.000 while the second was 0.981.

[chromatographer1]

20 micrograms of ethanol as vapor in 8mL of vial at 70°C.

And then you are sampling 2.0mL of vapor from this 8mL.

Adding to this you are spliting the sample 30:1.

I would think having SO MUCH alcohol in the vial you might have a saturation problem but after plotting the points I see a very consistent curve, especially over the last three points.

I suspect you have an overload on the column, or more likely as has been noted by posters previously on this thread:

you are changing the concentration by withdrawing so much of the headspace volume from the vial.

You should change the amount of alcohol in the vial, not change the volume of your injection to do a true linearity of your system.

I would also drop the volume of water in your vial to less than 250 microliters.

If you do this I believe your linearity should improve immensely.

You should be able to attain 0.999 linearity with ethanol.

best wishes,

Rod