I've no experience in GC and I get some problems.
I'm using headspace and MS Both on a 6890 (matrix DMSO,benzyl alcohol or water) and I "win" a pic at the beginning which doesn't allow to see solvents as methanol. Is there a solution to avoid the "mean" pic?
Is DMSO usually used in GC cause we have strange things since we use it as a matrix (principally on our FID)?
Thanks

I guess your stationnary phase is a HP-1 or HP-5 ?
If you have to use this kind of matrix and separate solvents try a stationnary phase like Poraplot Q HT or equivalent, MeOH and DMSO will be perfectly separated.
DMSO is known a weak volatility dilution solvent I don't recommend you to use it in another GC injection mode than Headspace.

Any column can easily separate MeOH and DMSO. Your DMSO contains some impurity. It can be MeOH or product of decomposition of DMSO.
To separate DMSO and MeOH you have to program the temperature of oven, for example, from 40 g to 200 g.

Sorry, my english's not good so you may not understand what I was asking. So, I use a static headspace (7694), a DB624, and either a mds or fid (6890+5973). My problems are that using fid, DMSO seems to damage my detector (get some spikes, changing noise, base line falls down quickly...)and using msd, we got what we call "the air pic" at the beginning and we can't see methanol which's under. In both cases, we get some problems to clean our transfert HS line. Hope it's clearer

DMSO can't damage fid. I guess fid can't be damaged by anything.
I think, using MS you can quantify methanol without separating it from air.

DMSO has impurities and should be distilled just before use in HS analysis. One can place a volume in a 22mL vial and subject it to repeated HS analysis to watch the coeluter with methanol decrease in size.
If redistilled it must be kept away from water and oxygen.
Dimethylsulfide and Dimethyl disulfide are commonly found impurities in DMSO, if I remember correctly.
DMSO itself does separate from methanol quite easily as noted in previous posts.
But it can react with residues left in transfer lines and valve sample loops, as well as any bare metal found in your HS system to give artifactual peaks.
To minimize this problem, use tubing that is fused silica coated or deactivated with some other process.
Two other impurities which can give coelution problems with methanol under certain conditions are formaldehyde and methyl formate.
It is suggested that you use a minumum of DMSO to dissolve your matrix and you determine that solubility at the temperature you plan to perform the HS analysis.
Use water or someother solvent to dilute your matrix if possible once dissolved into DMSO.
DMSO itself will not cause spikes or other detector problems but it will dissolve or release otherwise 'glued down' particles which can give you such artifacts.

Best wishes,

Rodney George
Senior Research and Development Scientist
Gas Separations Research
Supelco
595 North Harrison Road
Bellefonte, PA 16823

814-359-5737 voice
814-359-5459 fax
rgeorge@sial.com

Thanks for your precious help. I'm going to try all advices you said.

Thank you all for your advices, it'll help me