Diol column question

[sdnico]

I have been working on an older NP method for phosphtidylcholine using an ELSD detector. The method calls for a LiChrosorb Diol 60-5 250X4.6 column using a gradient with 57:43 Hexane:IPA (MP A) and 103:77:20 IPA:Hexane:H2Ow/NH4Ac (MP B).
Previous data for this assay (data approx. 10yrs old) shows that this always yields results within product specifications. My results are consistently 40-45% higher. I switched to a PVA-sil S-5 250X4.6 NP column which is used for other phosphtidylcholine products that we have here and the numbers magically dropped to within product specifications. It seems unlikely to me that the column would cause such a drastic change in the numbers since the sample is calculated against a standard curve.
Any ideas about the cause would be appreciated.

Thanks for the help.

[Mark Tracy]

I'm going to ask a bunch of random questions hoping to hit something...
- Do you get similar calibration curves with both columns?
- Does the PVA-sil column give you more peaks than the LiChrosorb?
- Do the peaks elute at the same gradient composition for both columns?
- Have you tried to optimize the ELSD parameters?
- Is the PC fraction percent of total area different for the two columns?

[sdnico]

Mark,
Thanks for the response.
In answer to your questions.....
The R-squared value is 0.999 for both curves
There are no additional peaks with the PVA-sil column
The ELSD has been optimized for the method. (The gain setting and injection volume are adjusted to ensure that the most concetrated std falls within a certain peak height range.)
The peaks do elute at different gradient compositions. The peak in the Diol column elutes at the begining of the gradient switch from A to B.
The PVA-sil peak elutes as mobile phase B has reached the end of it's 100% hold time.
As for the fraction percent, I don't have the numbers in front of me, but I don't recall them being significantly different.

Thanks again. If you see anything obvious please let me know.