I have had serious problems trying to use water as a diluent on an Agilent 5890 hooked up to an Agilent 5973 MSD, and I had to change the diluent. However, I think newer systems are amenable to using water as a sample solvent.

if you want to make a direct injection in the GC, you must use the innowax column.

We are using water as a solvent for GC(MS) frequently. In the past there was a research project on sorbic acid decomposition in aqueous acid environment. We applied direct liquid injection on a aquawax and aquawax DA column from Grace without problems. This kind of stationary phase is created for water injections. The only annoying thing is a huge water peak ( more a blob then a peak) around 5 min by 50°C

Most of the time we are using headspace . I have one system running a 624 type column and there seems to be no immediate problem with water (FID system).

If you have access to a headspace trap system it is even possible to remove the most of the water before injection leaving you a free choise of column

Thank you very much Jonathan Decenzi, Sandra Martín del Campo and Kris Wolfs..

@Kris Wolfs, can you please give the details of the column (dimentions) and conditions used.

Thanks in advance, Srikanth

Wath do you want? the sorbic acid direct injection approach or the headspace approach?

Just now, i discussed with my technical support and wrote a detailed mail to your official mail id. can you please look at it and let me know.

Have you think about SPME (solid phase micro extraction) followed by a HD (head space) and a PEG column. This has been proved to work well for determination of methanol in water, for instance. I think it would work with your analytes.

Why don't you try "purge and Trap" system to get these analytes into a sorbent and analyse by GC-FID after extracting them with a solvent?

Thermal desorption would be useful alternately.

According to this

http://www.chromatographyonline.com/lcgc/data/articlestandard/lcgcasia/082006/309391/article.pdf

injecting water into GC-MS is usually not a great idea. I'd use NMR instead, you don't even need D2O as solvent. Use water, add few drops of D2O (for lock) and suppress the H2O signal.

Thank you very much to all of you for the discussion and ideas. We have found a method from Kris Wolfs, which is suitable to our matrix.

@Kris, Thank you very much once again

It sounds like you already have a method but I just thought I would mention that some years ago I performed dose certification analysis of aqueous chloral hydrate solutions by GC/FID. Direct injections (1-4ul) of the the aqueous solution were made onto a Carbopak B / 5% Carbowax 20M packed column (6' x 2mm id) using N2 20ml/min, Inj temp 150C, Oven temp 110C isothermal, FID 180C. This was used on a regular basis over a couple of years without problems.

Hi Srikanth,

I'm analyzing literally the same compound you have and to my knowledge, NMR can work with this system but I'm also interested in trying a GC-MS method. Could you please send me the reference on GC-MS method? Currently I'm using GC-FID to analyze solely methanol on a Supelcowax 10 column.

In addition to Kris's methods (both of which are excellent) you can inject directly. Your sensitivity is limited to ppm range, but if you use a polar column you'll pull the formic acid out of the water peak. Contrary to some of the posts, you most certainly CAN inject water into a MS. If you use purge and trap you essentially do that with every injection.

Methanol and formic acid can react with each other to form methyl formate which could potentially complicate the chromatogram. Just something to keep in mind.

Carboxylic acids typically do not chromatograph well. I'm not sure how applicable it would be in your situation but I just found a method for formate analysis in J. of Analytical Tox, Vol. 31 Jul/Aug 2007 p 342. Here they convert formate to ethyl formate and analyze the headspace by GC/FID. GC/MS could also be used. Method modification may be required though in your case due to the methanol. If the methanol is high enough relative to the formic acid maybe you could could acidify a sample with H2SO4 and convert the formic acid to volatile methyl formate and analyze both it and the methanol, perhaps formaldehyde too. Aldehydes are reactive too though and all this would have to be evaluated.

I agree to Thomas and Mark. Take the right column (certainly not your standard DB1) use a dedicated machine adapt your Tem program and have fun. However, be aware that non volatile compounds like salts will quickly corrupt your sytem, if they are present.