Chromatography Forum

Peter, 1.1mL and 1.0mL will put the same concentration of gas in the headspace pretty much as my example(ethanol in water). The partition coeffcient dominates the denominator because it is so much larger than the phase ratio. So when you mass correct the 1.1mL you are creating a 10% error.

And the original inquiry was about volume tolerance and the effects of volume changes on sample analysis. What you have told her will lead her to generate what may be invalid results for her needs. She would be best served to perform robustness testing where she varies the weight or volume of the sample addition (preferably spiked samples with known concentrations) and analytically determine the effects of the volume change on her analysis.

jgarey

Karen is using much smaller volumes than you are discussing. This changes the dynamic, as I mentioned and Peter concurred.

She is using 200 microliters of a viscious liquid.

For your consideration:

If you performed a complete headspace evaporation of a tiny volume, 1 or 2 microliters, would the amount of volatile impurities in the headspace stay the same with either sampling?

In the same fashion, when the volume of sample is much smaller than the headspace volume the amount in the headspace is approximately linear to the volume of sample, or since the volume is difficult to dispense, the weight of the sample.

We are not on the same page in the experiment. Once we are I am sure you will agree with Peter.

best wishes,

Rod

Peter, 1.1mL and 1.0mL will put the same concentration of gas in the headspace pretty much as my example(ethanol in water). The partition coeffcient dominates the denominator because it is so much larger than the phase ratio. So when you mass correct the 1.1mL you are creating a 10% error.

And the original inquiry was about volume tolerance and the effects of volume changes on sample analysis. What you have told her will lead her to generate what may be invalid results for her needs. She would be best served to perform robustness testing where she varies the weight or volume of the sample addition (preferably spiked samples with known concentrations) and analytically determine the effects of the volume change on her analysis.

I am guilty of making an assumption - because Karen was worried about the effect of sample volume on headspace results I assumed that she had analytes for which volume would make a difference - which might have been completely wrong. However she did ask about adding an internal standard, and clearly then she has to know how much sample is in the vial, and the best way to determine that is by weighing, irrespective of the partition behaviour of the analytes.

What happens when Karen finds that her analysis is not robust to sample volume ? (and whether it is or not depends on whether any of the analytes partition strongly to the headspace) - what is you suggestion for how she can make the method robust ?.

Karen, forgive me if I have sent you down the wrong track (your last post leads me to think that this is not the case) - can you tell us what the matrix is (aqueous or non-polaor organic) and what the analytes are ?

Peter

I don't think the smaller volume matters. Using the ethanol in water example and using your small sample volume of 200uL. If you add 190uL and 210uL you will end up with the same concentration in the gas phase at equilibrium. Again because the high partition coefficient dominates, the 190 and 210 samples will transfer the same gas concentration of ethanol on column. But if you then correct for the mass of your sample, you're concentration calculations will be off by 5%. In this case she would be better off ignoring the mass difference, rather than correcting for it. If it were hexane in water she would be better of correcting for the mass. Again it depends on the matrix and analyte.

If she was trying to analyze surfactants for residual solvents like dioxane, then she is going to have higher partition coefficients. To offset this she may have to use larger sample volumes and longer equilibration times, sacrificing throughput for accuracy making the method more robust.

Karen, forgive me if I have sent you down the wrong track (your last post leads me to think that this is not the case) - can you tell us what the matrix is (aqueous or non-polaor organic) and what the analytes are ?

Peter

I can't give out the details but the solvent is viscous and has a BP of about 190 and the analyte about 120 and the headspace oven is at 70 (because i don't want to throw too much of the solvent into the headspace on a loop headspace sampler as I don't want it condensing in the vent). Both are fatty alcohols. The solvent is used to extract the analyte from aqueous media . That implies to me that not much (relatively speaking) will partition into the headspace.

Though as I said, the problem with dispensing should go away when I get the positive displacement pipette.

- Karen

Hi Karen

You have one alcohol dissolved in another, so the headspace concentration should be robust to sample volume (as jgarey has pointed out). I presume that the lighter one is a quite high concentrations if you are able to detect it easily in your pilot trials.

A further drawback of sticky samples is that they equilibriate slowly - the way around this is Rod's small samples - the thinner the film the quicker the equilibrium.

Peter

Good point Peter.

Please let us know Karen how your experiment progressed.

best wishes,

Rod

Not that I really want to get into this one but, practice back pipetting by calibrating your pipettes maybe. Or practise pipetting the volume you need to a constant weight on a balance.

If the sample is more viscous than say H3PO4 I'm not sure a micropipette is the best choice.

I do a lot of related compound and trace methods here and I find most problems can be traced to not using the pipette consistently...based on that, I always make the coop students calibrate them so they will learn to be consistent before they start helping me prepare anything!

You sound like an individual who likes to get correct answers,

and one who has run into human foibles a time or two.

best wishes,

Rod

You have one alcohol dissolved in another, so the headspace concentration should be robust to sample volume (as jgarey has pointed out). I presume that the lighter one is a quite high concentrations if you are able to detect it easily in your pilot trials.

In the % range.

A further drawback of sticky samples is that they equilibriate slowly - the way around this is Rod's small samples - the thinner the film the quicker the equilibrium.

That's why I'll be using 200uL.

Years ago I learned how critical pipetting was... In my first job back in the stone age, I had to do a lot of RIA ... results were all over the place if you did not not do it consistently!

- Karen

Not that I really want to get into this one but, practice back pipetting by calibrating your pipettes maybe. Or practise pipetting the volume you need to a constant weight on a balance.

The method will have multiple operators so I don't want it to require an "expert"!

If the sample is more viscous than say H3PO4 I'm not sure a micropipette is the best choice.

The positive displacement pipette (which came in last week) looks like it will work just fine , though i many not be able to get back to the method for awhile as other things have come up.

- Karen