Solvent/Column for PEG residual solvent determination

[sbkg32]

I am trying to develop one method to determine the amount of Methanol, Dichloromethane, methyl-tertbutylethylene, hexane, ethyl acetate, chloroform, and toluene in PEG.

I can get all but methanol using DMF as a solvent and a DB-624 column.

I also tried water as a solvent using 1.5 NaCl to release the solvents also with a DB-624 column. Methanol looked great, but the rest had very poor recovery results.

Any suggestions?

[krickos]

Hi

Before asking for more details about the headpsace parameters (assume chromatograph is fine and recovery is the issue), I have a different question as it seems like you have met "Mr sample matrix effect".

Do you use external standards?

PEG is quite likely a "strong" retaining matrix, hence the NaCl addition in water "kills" it for methanol that is a polar solvent but is "not enough" for the less polar solvents.

Unless you have not tried it already, try standard addtion.

Poor recovery in water is equal to low recoveries?

Cheers Chris

[mbicking]

I would repeat the experiment with only water as the diluent. Also, there is a USP method for residual solvents that can do many of these compounds, or similar solvents. The method itself is junk, but the data indicate that you should be able to get all of these by headspace.

Another option would be to try 10% DMSO/water as the diluent. We found that this seemed to help solubility of some organics and stabilized the systems.

[krickos]

Another option would be to try 10% DMSO/water as the diluent. We found that this seemed to help solubility of some organics and stabilized the systems.

Yes that is not a bad idea, however one need to be a bit careful when doing it, see figure 5 in this application note:
http://www2.shimadzu.com/applications/gc/g228.pdf

Also take not that DMF does not show the very same behaviour when diluted with water ie sensitivity increase does not seem to fall back after 50% as with DMSO.

The catch with % mixtures of DMF/DMSO in water is the expotential sensitivity increase you get for most solvents up to a 50% mixture. This in turn makes the preparation of the mixture important (metric cylinders is not that accurate ).
As described in one of the USP methods, it may be better depending on % mixture from a reproducibility point of view to prepare stock standards in DMF/DMSO (longer stability) then dilute with water, accurately add 1ml of DMSO/DMF to your vials then add water based sample/standard solutions.

Even if this helps some with sensitivity and stabilisation of potential water vapour issues, it does not necessary solve the sample matrix issues. Those may still require addition of salts like NaCl or sodiumsulphate alternatively the use of the standard addition technique.

uhh got a bit longer than I thought but felt necessary.

[mbicking]

krickos:
Our work with 10% DMSO was based on the solubility and stability of the standards. As designed by USP, some of the solutions are not stable. The DMSO seems to help this.

We did not see any signficant changes in response, but precision was much better. The system met the sensitivity requirements easily.