Solvent peak shifts on GC-headspace analysis

[osp001]

Run into an interesting problem with headspace analysis, looking for solvents in extracts. See attached image.



The top (Data 1) is an example of our analyte, which is extracted with ethanol. The bottom (Data 2) is pure ethanol. Note the peak of interest (around 1.7 minutes) is shifted. A close-up of this shift is in the second panel.

Ethanol elutes at 1.791 minutes with great repeatability. Similarly, the peak in Data 1 (1.650 minutes) reliably elutes at that time.

Nothing tricky yet.

What's weird is that when we spike a sample of the analyte with ethanol, the peak at 1.650 minutes gets larger, and there is never a peak at 1.791 minutes. We know the peak at 1.650 is ethanol, but it should be coming off at 1.791. So what's happening?

Samples are run on a 30 m x 0.25 mm x 1.4 µm Rxi-624Sil MS column, headspace extracted after 20 minute equilibration at 140C.

[Peter Apps]

Tricky - the peaks in the enlargement look a bit overloaded perhaps, but that should produce the opposite shift in retention to what you are seeing. Is there any water in the sample ? - water on the column easily shifts the retention of early eluting very polar compounds.

Peter

[osp001]

I didn't know that about the water shift- that could well be it, as even if the extraction were performed with anhydrous ethanol (it's not), the plant material is not absolutely dry.

Does that mean anything in terms of quantitation? Would peak area be affected?

[Peter Apps]

Quantitation will not be affected. If the shift is really a rpoblem you could try adding a little bit of water to all the samples and standards so that the extra retention is always present. Since you have more than enough separation of ethanol from the other components I would say that the shift is not significant.

Peter

[osp001]

The odd thing is that I ran ethanol + water for my calibration, from 100,000 ppm ethanol down to 1 ppm ethanol in orders of magnitude, and saw no shift whatsoever in those standards- all came out at 1.791 minutes, +/- 0.001.

[Peter Apps]

The odd thing is that I ran ethanol + water for my calibration, from 100,000 ppm ethanol down to 1 ppm ethanol in orders of magnitude, and saw no shift whatsoever in those standards- all came out at 1.791 minutes, +/- 0.001.

That's because there is always enough water to increase the retention.

If you are calibrating with ethanol in water then you need to dissolve your samples in water, otherwise you will have very different partitioning of ethanol to the headspace between samples and standards.

Peter

[thohry]

The odd thing is that I ran ethanol + water for my calibration, from 100,000 ppm ethanol down to 1 ppm ethanol in orders of magnitude, and saw no shift whatsoever in those standards- all came out at 1.791 minutes, +/- 0.001.

That's because there is always enough water to increase the retention.

If you are calibrating with ethanol in water then you need to dissolve your samples in water, otherwise you will have very different partitioning of ethanol to the headspace between samples and standards.


Peter

Thanks Peter. I sometimes have this kind of problem analyzing alcohol.