Method Development Advice

[jdyer51287]

Hello,

I am currently trying to develop a method to separate levoglucosenone, HMF anf levoglucosan in bio-oil samples.

I am using a Zorbax SB-C18 (4.6x250mm, 5µm) column set at 40°C. I initially got the following chromatogram which shows adequate resolution of the peaks.



However, these bio-oil samples also contain a number of other sugars that co-elute with levoglucosan (I ran a lot more sugars than are included on the chromatogram). I realise that I will not get good sugar separation using a C18 column, I would just like to get the levoglucosan peak more resolved from the other sugars. I am happy for all the other sugars to elute as one peak.

By slowing down the flow rate from 1 mL/min to 0.6 mL/min I was able to move the levoglucosan peak so that it was not directly co-eluting with the other sugars but it is still not in any way resolved! It also has the detrimental effect of broadening my HMF and levoglucosenone peaks!



As you can see so far I have only run isocratic. I believe I might be able to resolve levoglucosan if I introduce a gradient at the beginning of the run, however I am not entirely sure what would work best. I was going to start very basic with something like:

0 - 85:15 H2O:MeOH
5 - 80:20
10 - 80:20
15 - 85:15

I don't know whether this would work or I'd be doing completely the wrong thing!

Any help would be greatly appreciated.

Thanks

[tom jupille]

First thing: Levoglucosan is essentially unretained. It is eluting at the dead time of the column ("t0"). You can try decreasing the methanol to 5%, but then you will probably need a gradient to get the HMF and levoglucosenone to elute at a reasonable time.

Second thing: the peak shapes show definite problem. How old is that column?

General advice: get a copy of Mike Dong's "Modern HPLC for Practicing Scientists" book (here's a link to it on Amazon:
http://tinyurl.com/7dnm8dw ); that will give you some insight into why I made the first two comments.

[jdyer51287]

Whoops, I completely forgot about dead volumes! It's been a fair while since I did any method development.

The column is very old and has not been treated very carefully but as it is only a research project there is no budget for a new column. That kind of peak shape should just about be acceptable for us for now.

Thanks for the link, I'll take a look.