Perhaps a mixed ion-exchange resin is what you need.

I agree wit Michael, but you might be not easy to find the right conditions (to optimize the pH for binding anions and cations may be tricky)

You can check out page 16 in the following document:

http://www.sigmaaldrich.com/content/dam/sigma-aldrich/docs/Supelco/General_Information/1/solid-phase-extraction-products-FEB.pdf

Did you consider to use SPE cartridges which actually retain the carbohydrates - there are quite nice graphite-based pipette tips available (i.e. SUPELCO 54227-U), which work well for removing salts prior to the MS analysis - we use them all the time...

Good luck!

I just realised that, because my samples are very alkaline (10-12.3) it's possible that many of my sugars are charged, which would difficult their separation from other ions such as phosphates and calcium. I also would like to minimize the manipulation of the sample (pH changes, etc) before derivatisation as I want those samples to be representative of the experimental conditions.

What if I take another approach: first I derivatise (silylate) my monosacharides using TMS and afterwards I purify the samples from calcium and phosphate. Now, once silylated the sugars are not supposed to be that polar anymore, right? Then maybe adding a liquid-liquid extraction step would help separate the water-soluble calcium/phosphate ions from the non-polar silylated sugars. The solvent of my derivatisation is pyridine, which is quite polar, but I could evaporate it and resuspend it in DCM and then do the LLE by adding water, and then getting the DCM phase back.

If this is not possible for any reason, I guess I could also try a SPE purification after derivatisation if I use a SPE matrix that would adsorb my silylated sugars and let phosphates/calcium run through (or the opposite). I don't know which approach could yield faster and better results though. What do you think?

Thanks!