Headspace Sample volumen tollerance

[Karen01]

I need to measure an analyte that is in a somewhat viscous solvent that boils at a higher temp. The sample will be received that way.

A quick test shows that I can see the levels of the analyte I need to via headspace FID without putting too much of the solvent into the headspace (I'm using a loop Headspace sampler) . What I worry about is dispensing the standards and samples reproducibly.

I have always been able to measure volumes sample and standard solutions into headspace accurately.
The solvent is viscous enough that I worry about reproducibility of the pippetting and adding by weight is hard to get precise.

How critical is having the same volumes fro headspace?

Thanks
- Karen

[Peter Apps]

adding by weight is hard to get precise.


Thanks
- Karen

You should be able to weigh to 1 mg in 1 gram very quickly, that's equivalent to 1 ul in 1 ml. I doubt that you can be that precise volumetrically, even with water.

Peter

[carlos.teixeira]

Dear Karen,

I would add to the Peter's comment, the efficiency of your HSGC analysis may depend not only on the sample volume (in your case, solvent volume) but HSGC flask, too. Your result may be affected to both.

Hoping the best,

Carlos de Souza Teixeira
teixeiracs@yahoo.com
http://br.linkedin.com/pub/carlos-de-so ... 15/7b0/125

[Karen01]

adding by weight is hard to get precise.


Thanks
- Karen

You should be able to weigh to 1 mg in 1 gram very quickly, that's equivalent to 1 ul in 1 ml. I doubt that you can be that precise volumetrically, even with water.

Peter

Thanks but I think you misunderstood.
It's not making up the standards that is the issue. it's putting the same volume of samples and standards into the headspace vials. In the initial experiment I put about 150 mg (200 ULof a standard vial with an air displacement pipettor) a 10 mL headspace vial. I don't want the smples to take too long to equilibrate.



I have always matched sample and standard volumes very closely for external calibration for headspace. I guess I'm asking how critical that really is when using headspace.

- Karen

[Bigbear]

I do blood ethanol via headspace and some samples are quite viscous. I use a positive displacement pipette with wide opening tips. Seem to work quite well. In the past I have even clipped the pipette tips to give a wider opening.
If you are very concerned about sample volume you could try adding an internal standard to your samples.
What concentrations are you working with in your samples? If high you could consider diluting the samples prior to pipetting.

[chromatographer1]

Karen,

Assuming you are using std vials, 12-20mL in volume:

Your answers will vary with the variance of the volume of the sample sampled.

If you transfer 180 microliters of sample for one vial and 220 microliters of sample in a second vial, your reproducible sample will differ by 20%

The first might give you 0.4% and the second will give you 0.5% for the volatile you are measuring.

Is this the accuracy you desire?

But if you can transfer 198 and 202 microliters in two attempts, then your answers will vary by 2%.

This answer depends on the sample volume being much smaller than the vial volume.

best wishes,

Rod

[Karen01]

Thanks all... I'll just have to find a way to add sample precisely and accurately.

[Peter Apps]

Hi Karen

Using Rod's unquestionably effective small volume in large vial procedure, you can correct the result according to the actual amount of the sample, which is easily, accurately and precisely measured by weight.

You add an amount of sample or standard (as close as you can conveniently get it to the nominal 200 ul target) and weigh what you actually added. Then you divide the peak area for the analyte by the actual weight.

If you are in a regulated lab it might be worthwhile demonstrating that peak area is linear vs sample weight over the range that you expect to be adding.

Peter

[walter]

You can perform a method optimization routine by varying the equilibration time to determine the optimum time based on analyte response. If you double the sample weight you do not always double the response or the equilibration time depending on the matrix and analyte. We routinely use 0.5 to 1.o gm samples in 10 to 20 ml vials.

[jgarey]

I think everyone is misunderstanding the question. For example, if one adds 0.5mL and 0.6mL of sample to separate vials the difference in response will be caused by both the change in sample quantity and in headspace volume. So you can't really just mass correct your results. At 198uL and 202uL the volume difference is negligible so you probably can. But, I think she was more curious as to how large the impact of changing the headspace volume was. Which just depends on the samples and analytes of interest.

[Karen01]

FWIW I'm going to try to solve the issue by getting a positive displacement pipettor which is on order. Hopefully that will solve the dispensing issue for the viscous liquid reproducibly.

- Karen

[Peter Apps]

I think everyone is misunderstanding the question. For example, if one adds 0.5mL and 0.6mL of sample to separate vials the difference in response will be caused by both the change in sample quantity and in headspace volume. So you can't really just mass correct your results. At 198uL and 202uL the volume difference is negligible so you probably can. But, I think she was more curious as to how large the impact of changing the headspace volume was. Which just depends on the samples and analytes of interest.

You can indeed "just mass correct your results", how do I know, becuase I have done exactly this. A little simple arithmetic will show that for a 0.5 ml sample in a 20 ml vial, any change to the sample volume has 1/40th as much relative effect on the headspace volume. And since headspace analysis measures the concentration (strictly mole fraction) in the headspace the only effect of headspace volume (presuming as in this scenario that the liquid phase is not depleted of analyte) is a result of pressure changes as the headspace sample is withdrawn.

Peter

[jgarey]

Sorry, but that only applies to analytes with low partition coefficients. We can compare xylene and ethanol in water at 50C.
Concentration of analyte in gas phase= Concentration of analyte in matrix / (Phase ratio + partition coefficient)
K=1150 for ethanol under those conditions. So assume you have 100ug/mL of ethanol in water solution. You add 1 mL to a 22mL vial and equilibrate using static headspace with a loop.
Cg = 100ug/mL / (21/1)+1150 = 0.085 ug/mL
Now if you add 2mL
Cg = 100ug/mL / (20/2)+1150 = 0.086 ug/mL

Compare to xylene with K=1.3

Cg = 100ug/mL / (21/1)+1.3 = 4.48ug/mL
Cg = 100ug/mL / (20/2)+1.3 = 8.85ug/mL

In samples with low partition coefficients (K=Cg/Cm) under the method parameters you can mass correct with possibly non-significant error. But, like I said before, it depends.

The volume change can also create an issue in headspace systems that have a vial pressurization step (like Tekmar HT3). The different volumes will create different static vial pressures (as mentioned by Peter). So the vial pressurization step will add different amounts of gas to bring the vial pressure to the setpoint.

[Peter Apps]

The effect of partition is, of course, very important. But I would venture that a doubling of sample volume is way outside the kind of dispensing error that it would be sensible to try to correct for. Try the same calculations with 1 ml and 1.1 ml.

And to get back to the original enquiry about how best to work with viscous samples that are difficult to dispense volumetrically; no matter what the analyte, correcting for sample mass will reduce the effect of dispensing on the final result and possibly bring repeatability RSD down to the required level without the need for very elaborate dispensing procedures, which might themselves introduce errors, especially with analytes that partition strongly to the headspace.

Peter

[chromatographer1]

Peter has made excellent points.

Handling the samples for extended periods of time, using a partial vacuum to dispense a portion from the bulk can change the result in the loss of highly volatile components and should be avoided.

This is why he is correct that a quick and simple weighing of the portion taken should correct the HS measurement more accurately than the endeavor to transfer an exact volume.

Good luck with your work Karen.

Rod