Matrix Effect in Headspace

[Jdougherty88]

I have been trying, unsuccessfully, to quantify 5 residual solvents using headspace chromatography that may reside in a polymer that I have manufactured. At first I used an external standard method, but noticed that the areas of the standards (just analytes in solvent) were much different for 2 of the 5 analytes than the areas of the samples (polymer + spiked analytes). The other 3 analytes were consistent. I have raised the incubation temperature to 140C and have tried increasing the incubation time to 1 hour, but the 2 troublesome analytes still are not giving consistent results. I tried to analyze the samples by using the standard addition method, but still I do not have consistent results. Can anyone make any suggestions on this? Thank you!

[rb6banjo]

Tough problem. Partitioning behavior of volatiles in plastics are never - in my experience - the same as in some solvent. There are many ways to attack this problem. If you can provide more information as to the identity/nature of your polymer it would be helpful. To reliably add analytes to a polymer, you must be warmer than the glass-transition temperature for the polymer. Even then, it can still take a long time. Can you share with us, the identity of the analytes?

I work with highly plasticized (~40% plasticizer) pvc and since it has so much plasticizer, the added analytes partition into the plastic fairly quickly (on the order of hours at 70 °C). It behaves like a very viscous liquid. For materials like polystyrene it can take much longer.

[LiVD]

You might want to take a look at multiple headspace extraction.

[Jdougherty88]

I am attempting to detect Methanol, Ethyl Ether, Ethylenediamine, Toluene, and 2-Ethyl-2-Oxazoline in Poly(2-Ethyl-2-Oxazoline). The Methanol, Ethyl Ether, and Toluene all give me consistent results. The Ethylenediamine and 2-Ethyl-2-Oxazoline, however, give very poor results when compared to the standards.

[rb6banjo]

Darn. Those are the 2 in that list that I'd expect to the be the most misbehaved. The problem may not be the partitioning of the analytes out of the polymer. They could be out of control because they are sticking to the sample path from the vial to the injector.

What are the requirements for the detection limit of your method? If they are not too low, you might try dissolving the polymer in a solvent. This usually makes the detection limit higher BUT it can help to eliminate matrix problems caused by inadequate partitioning behavior between the sample and the headspace.

If the problem is in your sample path, it's much more difficult to overcome. Perhaps some type of direct injection technique (gas-tight syringe, or SPME)? That will at least eliminate a long transfer line from the equation. I have had success with this approach when analyzing for materials with that are reactive (carboxylic acids, amines, etc.).

[Jdougherty88]

The polymer and the standards are already in N,N-Dimethylacetamide (DMAc) solvent. I use a gas tight syringe with the CTC Leap Autosampler.
Example of my standard/sample preparation:
1. Prepare a 10 mL standard containing 60 ppm of each analyte (methanol, ethyl ether, ethylenediamine, toluene, 2-ethyl-2-oxazoline) in DMAc.
2. Add 50 mg of polymer sample to headspace vial. Transfer 100 uL of 60 ppm standard to the headspace vial to dissolve the polymer.
3. Transfer 100 uL of 60 ppm standard to a 2nd headspace vial.
4. Analyze both headspace vials under same conditions.

A Generalization of Results:
Standard Polymer Sample
Analyte Area Analyte Area
Methanol 80000 Methanol 80000
Ethyl Ether 100000 Ethyl Ether 100000
Ethylenediamine 30000 Ethylenediamine 21000
Toluene 110000 Toluene 110000
2-Ethyl-2-Oxazoline 50000 2-Ethyl-2-Oxazoline 23000

Same Solution, Same Conditions, Different Results. I might have to try MHE to SPME.

[Peter Apps]

What you are seeing is almost certainly the effect of the dissolved polymer on the partition of the analytes between headspace and liquid. This is to be expected, and as long as it is repeatable it should not present any problems.

Peter

[Jdougherty88]

Ok thank you everyone! In order to quantify the analytes will I have to use the standard addition method?