Problem with Octanoic Acid peak.

[dogmatta]

Hello all,

I am trying to run a test for impurities using octanoic acid as an internal standard. I keep getting an RSD failure because the octanoic acid peak area keeps falling each injection. The method has been validated and run before but this problem keeps cropping up. I am wondering if anyone has encountered this and if they have any fixes that would not require a revalidation. I am using a liquid inject method on a shimadzu gc 2010 and a liner with deactivated wool. The column is an agilent EC-1000 (30 m 0.32 mm 0.25 micrometer phase thickness) with a 5 m guard column. The diluent for the method is denatured ethanol. Here are the method parameters:

Injection Port Temperature: 220°C
Split Ratio: 10:1
Detector Temperature: 250°C
Carrier Gas: Helium
Carrier Pressure: 4.4 psi (~30 kPa)
Injection Volume: 1 µL
Detector Gases: Air: 400 mL/min
Hydrogen: 47 mL/min
Run Time: 50 minutes
Column Oven Program: Initial Temperature: 130°C
Initial Time: 3.5 min
Rate 1: 3°C/min
Temperature 2: 200°C
Hold Time: 0 min
Rate 2: 25°C/min
Final Temperature: 225°C
Hold Time: 22.17 min
Column Pressure Program: Initial Pressure: 4.4 psi
Initial Time: 25.0 min
Rate 1: 1.5 psi/min
Final Pressure: 11.6 psi
Hold Time: 20.0 min

Any ideas would be greatly appreciated.

[Peter Apps]

My first guess would be that the acid and the solvent are reacting to make ethyl octanoate. Check your chromatograms for a peak that is getting larger with each injection - this is the ester that used to be your acid.

The method might have been validated using ethanol that had just enough water in it to inhibit the esterification.

Peter

[dogmatta]

Thank you, Peter. Unfortunately, the last time this was tested we had the same problem with ethanol that was not denatured. Using denatured ethanol was thought to be the fix for the method, but I am finding that is not the case.

[Johnny Rod]

Denatured just means it has some methanol or t-butanol or something else in it, it can still work as ethanol. If Peter is right, try adding a little water to your next standard and see if the same thing happens. Octanoic acid seems an odd choice for a GC standard but I guess you've had success with this before. Have you changed the guard column?

[Consumer Products Guy]

And what is your analyte? Octanoic acid (and other similar acids) typically are somewhat difficult for which to obtain good peaks. We derivatize soaps and fatty acids to methyl esters for analysis for just that reason.

What I'm saying is that octanoic acid may not be the best choice for internal standard. Especially when dissolved in an alcohol and there's heat from the GC inlet.

Fatty acid + alcohol + heat = esters, as the others explained above. And additional acidity will just help catalyze the reaction.

[dogmatta]

Thanks to all for the quick responses. The analyte is valproic acid. The column and guard column have been changed multiple times. I am trying to avoid developing a new method, but it is sounding like that is the way I am headed.

[Peter Apps]

If you can possibly avoid it, do not use denatured ethanol - which contains impurities that can only complicate matters and which are probably not consistent from batch to batch. Check that the original method did not use 96% ethanol - this is what you get if you do not specify "absolute" ethanol.

Peter