Irreproducible chromatogram of soybean oil

[jschlechte]

Here's the problem:

Day 1 - HPLC system is up and running with 35/65 ACN/Acetone. Flushed for a couple hours and ran a sample of 1% soybean oil (SBO) dissolved in the mobile phase. Being more of a supplemental advisor of the operations (I have some HPLC experience) than the operator (a graduate student never used HPLC before who is working with someone who has limited HPLC experience as well) the UV was operating @ 254 nm (not good for what we wanted) and the DRI cell hadn't been purged.

Day 2 - After switching to 30/70 ACN/Acetone for better solubility and purging the DRI cell along with changing the UV wavelength to 209 nm (for ester in SBO) we obtained a decent chromatogram. The peaks were not as sharp as we would've liked and tailed some, but they resembled what had been done a couple years earlier on the same column. Both UV and DRI signals were in good agreement.

Day 3 - Switched to 70/30 ACN/Acetone ran a more polar small molecule (diester with 2 dioxane rings) that eluted with the solvent peak. Tried a sample of SBO and got a peak with/just after the solvent peak.

Day 4 - Switched back to 30/70 ACN/Acetone and have been unable to reproduce the SBO chromatogram. Some separation, but absolutely awful peak shapes and very small signals (0.06 AU now as opposed to 0.8 AU before) and essentially no signal from the DRI.

Days 5/6 - Flushed system with large amount of mobile phase, switched to strong solvent - THF and flushed. Went back to mobile phase and same problem, what peaks we do see are very small and distorted.

Thoughts I've had but not yet tried - back flushing the column and cleaning/testing the detectors for signal.

We are using a Novopak C18 300x3.9 mm column.

[Kostas Petritis]

Are you really using that much acetone with UV detection (acetone absorbs in UV)?

And your mobile phase seems very apolar to me to be used with a C18 (but not impossible)...?

[jschlechte]

Are you really using that much acetone with UV detection (acetone absorbs in UV)?

And your mobile phase seems very apolar to me to be used with a C18 (but not impossible)...?

Yes, 30/70 ACN/Acetone gives relatively good resolution, we were curious about lower acetone content, but that's when things started to go wrong.
As for the UV absorbance, the UV-vis spectrometer gives a peak inspite of the acetone (yes, I know it isn't ideal, but that's the only wavelength that will work).

Available info just on the web gives reference to acetone/ACN and dichloromethane/ACN as mobile phases for vegetable oils.

[Consumer Products Guy]

We use ACN-ethanol on RP column with ELSD detection for triglycerides like this. Yes ethanol.

[jschlechte]

We back flushed the column for a couple hours and the peak shape looks slightly better, but the overall signal is still a small fraction of what it should be. However, the solvent peaks are still strong, so it doesn't look like the detectors are the problem.

Is it time for a new column? Any suggestions on columns for vegetable oils? We also modify the oils (epoxidize and ring open to form hydroxyl groups). Is there a column that would be relatively good at both types of separations?

[Uwe Neue]

The suggestion to use acetonitrile-ethanol mixtures is a reasonable one, definitely better than acetonitrile/acetone. If ethanol is too expensive for you, you may bet reasonable results by substituting the ethanol with methanol or isopropanol or n-propanol.

[Cliff Mitchell]

Or you could keep costs low and get a few bottles of Everclear 190 proof. Don't laugh.....