Page 1 of 1

Ethylene Glycol

Posted: Mon Jul 12, 2010 2:38 pm
by Yuhuan Tang
Does anyone know good GC residual solvent conditions for ethylene glycol? I have been having trouble to get better sensitivity and reporoducibility.

Thanks,

Posted: Mon Jul 12, 2010 5:45 pm
by vballchemist
Is this direct aqueous injection looking for Ethylene Glycol. I use a NUKOL column and report 50ug/l for Ethylene and Propylene Glycol in water samples.

traces of ethylene glycol in water

Posted: Tue Oct 12, 2010 12:03 pm
by Palki
Hi vballchemist,
I have a problem with analysis you mentioned in your recent post.
We have a water samples with traces of ethylene glycol (ppm amounts).
But our GC system is not able make an good reproducibility and sensitivity.
We inject samples onto cappilary column (Innowax) through classical liner.
There is a problem with water evaporation volume...

You mentioned a NUKOL column but you did not mention type of injection and so on...please if it is possible can you send me some details?...
We tried packed column, direct injection, unusual liners...no solution of problem.
thanks in advance

My email: ppalkovic76@gmail.com

Peter

Posted: Wed Oct 13, 2010 1:32 am
by Consumer Products Guy
USP glycerin monograph lists impurities test for ethylene glycol (and diethylene glycol) on a -624 type column. We've used that for EG.

Posted: Wed Oct 13, 2010 10:38 am
by cjm
I test for ethylene glycol and propylene glycol in water using direct aqueous injection with a detection limit to 1ppm on an hp6890 using a wax column and have always had difficulty with reproducibility. The solution was to change injection to 50%water/50%methanol. Reproducibility should improve, not to mention peak shape! Also a higher split ratio helps with the water volume expansion problem, but at expense of sensitivity. I use straight liner with wool, split 5:1 H2 carrier on stabilwax 30m 0.32mm 0.5um

I still have some issues with reproducibility, but I have eliminated causes from injection. The only issue now is with the phase itself. Trimming loops regularly solves this, but can get expensive.