Negative Peak in Normal Phase HPLC during Method Transfer

[xliang]

Hi all,

Hope you can help me resolve this mystery.

I have developed a normal phase method to separate an amine with two chiral centers. Below is a brief summary of the conditions:

HPLC Column: Chiralpak AD-H chiral column, 4.6 mm x 250 mm, 5 µm
HPLC Column Temperature: 40oC
HPLC Mobile Phase: 80/20 (0.05%MSA in hexane/ethanol, 200 proof, v/v)
Flow Rate: 1 mL/min
Detector Wavelength: 210 nm
Injection Volume: 5 µL
Run Time: 15 min
Sample Diluent: Ethanol
Void Time: 3-4 min.

The method separates four isomersThe baseline and separation is good during the method development and qulification at our site using both old and brand new columns.

However when this method was transferred to an outside lab, a significant negative peak around 5.5 min, interfering with the main peak analysis. In addition, the separation of the last two isomers are not as good as the separation we observed. Discussion between the two labs revealed that there was some incompatible issues during system conversion, such as the reverse phased injector was purged with normal phase mobile phase without IPA wash in between. The problem was presumably corrected since, but withough improvement of the baseline and separation. It's also interesting to notice that the injection of mobile phase produces a small positive peak at the same retention time.

It's OK to switch the sample diluent to the mobile phase to eliminate the negative peak, but the separation still can't be reproduced. We are really puzzled since we never detect it before. Any clues will be greatly appreciated.

[Noser222]

Maybe their ethanol is contaminated with water, if they are observing a much more prominent negative peak.

[xliang]

Thanks for your quick response!

It might be water. However, after the incompitable switch was observed, I asked them to flush the system with IPA, water, then IPA, then back to normal system and the problem is still there. Do you think water could still be there?

One thing I forgot to mention is they observed positive peak when injecting hexane itself, and negative peak when injecting ethanol and smaller positive with mobile phase.

Thanks!

[lupetto]

in case you've not done it already, i would run the test mixture for the column of lab#2, just to check whether the stationary phase is still ok.

[Noser222]

What I'm referring to has nothing to do with the HPLC system. Are they using the same grade of anhydrous ethanol? Is their bottle of ethanol older? If so, it may no longer be anhydrous.

[Bruce Hamilton]

I'd also worry about the issue of the negative peak affecting the analyte peak. If the peak is due to a component in the solvent, maybe the method needs more work to move the analyte peak from the solvent.

I'd also ensure that you are both using the same grades of all the solvent components ( including MSA ), and closely look at their absorbance at 210nm.

I'd want to ensure that any IPA used for flushing is also suitable ( anhydrous, low UV HPLC grade ), as it could be the IPA that is affecting the column performance.

I'd want to know how long they are conditioning the column, how they are storing the the column,and monitoring whether the resolution changes with injections during along sequence.

Please keep having fun,

Bruce Hamilton