Sample preparation

[Kristalin]

Hallo!
I am tring to analyse VOC content in the paits and varnishes by GC/FID. I have problem with sample preparation, i weigh the paint, dissolve it in methanol, vortex and centrifuge then, but my results are much smaller than actual. Does anybody have experience with that kind of sample preparation?
Thank you!

[MikeD]

Kristalin

We don't do this measurement every day, but when we do we weigh the methanol, then we quickly add the paint, varnish, adhesive etc. and weigh again. We try to keep evaporation of VOCs to the minimum possible at this stage.

[Kristalin]

Thanks for such a prompt reply. What is the VOC content in your samples, is it advisible to use GC when VOC (heavy gasoline) content in the paint is around 50 -75 %. Or it is better to use weighting method? But then we still need to analyse the compounds.

[MikeD]

I am not sure I understand you. We do not have a weighing method without GC.

We always use gravimetry when preparing samples for GC analysis. It applies to almost everything we do, not just volatile solvents in products.

Where we may differ from you is that usually we would not analyze the methanol dilution directly by liquid injection onto the column. We may vaporize a small known amount of the methanol dilution onto a sorbent trap and thermally desorb it into a GC. The non-volatile fraction is collected on inert disposable material such as glass wool. Semi-volatiles should pass through, but it depends on the definition of semi-volatile. Hydrocarbons up to C16 are volatiles in our definition.

[Peter Apps]

Hi Kristalin

How are you calibrating ?

What is the ratio of paint to methanol ?

Please describe your GC method.

Peter

[Kristalin]

Hi Peter!
My college did the sample preparation, she took droplet of paint and weighted it, added 6 ml methanol (diluted more later). It was done according to ISO standard 11890, although it suggests that VOC content should be less than 15 % for this method. Otherwise you should use difference method. But later we would like to use gc/ms for compound identification.
As our everyday work is to analyse diesel fuel index, we used the same conditions, and calibration, although we know that VOC in the paint is heavy gasoline, we calculated the results with internal standard c32.
According the certificate VOC content is 50 g/l we got just 20.

[Peter Apps]

Hi Kristalin

You might be losing some of the volatiles during sample preparation, especially during the weighing of the original sample. Are you doing everything, including the weghing, in closed containers ?

Depending on the paint, the solvent could be anything from C6 to C16 aliphatic hydrocarbons, with some toluene, xylenes and ethyl benzene, and some volatile esters or alcohols etc. All these are much more volatile than the C32 that you are using as an internal standard, and so you probably have discrimination in the GC inlet, but that can only be a guess because you have not given the GC conditions.

If your samples contain C32 the internal standard peak will be enhanced and your results will be biased low.

What result do you get if you use external standards ?

Peter

[MikeD]

Kristalin,

I don't use ISO 11890 and never read it before. Since my company has a subscription I downloaded parts 1 and 2 today and these are my observations.

Did you weigh 1-3 grams as in sample preparation clause 9.6 of 11890-2 ? You mentioned a drop of paint which sounds less than 1 gram.

Clause 7.1 says:
The internal standard shall be a compound which is not present in the sample and is completely separated from the other components in the chromatogram. It shall be inert with respect to the sample constituents, stable in the required temperature range, and of known purity. Compounds such as isobutanol and diethylene glycol dimethyl ether
have been found suitable.

I suppose C32 meets the letter of this requirement but it is not what the standard drafting committee would have expected. Try another internal standard that better matches the volatile fraction.