I'm trying to analyse alkonolamines with GC. I use drivatization with MSTFA and the column I have used was DB5 and is now BPX5. With the DB5 it was o.k. for a while but than the peak form became horrible, with a hill in front of the peak. Heating out, cutting the column 60 cm and cleaning the injector didn't help, but a new column have brought good results. But - after some weeks - the same peak-form. Now I wanted to try a BPX5 with fore-column (or guard column, I don't know the english word), but the peaks are as bad as with the old DB5. I don't understand this. Has anyone an idea?

We use either DB-1 or DB-5 types for ethanolamines through triethanolamines. Getting these to fully derivatize the hydrogens on the nitrogens can be a little tricky. But, sorry, that information is considered proprietary.

Beate,
Not sure why you want to derivatize the analytes. Restek shows analysis on a Rtx-35 Amine column without derivatizing. They show mono-, di-, and triethanolamine eluting with good peak shape. Check them out at www.restekcorp.com

Regards,
Mark

Guard columns can be problematic when running very active analytes like these. This is because you can consider the deactivation layer on the guard column to be very "thin" and thus you may still get some interaction with uncapped silanol groups on the guard tubing (no deact procedure is 100% efficient).

The description you give of the problem on the DB-5 makes me think that you were gradually contaminating the DB-5 with material from your injected samples and that was causing the loss of performance over time. This, of course, assumes that your derivatization procedure is good and that your deriv analytes remain stable in solution (ie. that you didn't have a problem with standard stability). Even very trace amounts of contaminant material (inorganic or organic) from injected standards or samples may cause adsorptive losses or perhaps decomposition of these analytes in your system. My suspicion is that you may end up seeing this same problem occur again on another manufacturer's column -- no column manufacturer can make a column that is "immune" to contaminants, that's impossible.

Did you ever try trimming more from the front of the DB-5 improved performance? Some contamination problems require trimming of up to 0.5 to 1.0 meters (or more!) before the material has been removed from the flow pathway. If you're doing temp programming then you can also try reversing the direction of the column in the system (install detector end in inlet and vice versa) for troubleshooting -- if response is better with "reversed" column then it's a sure sign that the front of your column is just very contaminated or damaged in some way.

I've had pretty good luck running monoethanolamine and triethanolamine on an HP-5ms column down to low levels underivatized. Of course, all other components in the system need to be very inert as well. My experience was with cool-on-column injection to eliminate adsorptive losses in the inlet and just evaluate column performance. Of course, your column can be very inert but performance will degrade rapidly if you're depositing material from your injected samples in the system that interacts with your target analytes.

Best regards,
Jason Ellis
GC Column Tech Support
Agilent Technologies

Check this link, we have developed this method for one of our clients. They are analizing ethanolamines in water matrix (matrix also have a lot of inorganic salts, impurities, etc.). Good peak shape and retention

http://www.allsep.com/makeChr.php?chr=Chr_041

Sorry forgot to ad that this is HPLC method and not GC (no derivatization)