The first thing I would try is just using different mobile phase combinations to see if you can get separation. What are you currently using?
Last ditch I would start looking at different crystallization conditions, hexane/ether and benzene/cyclohexane are the two main ones I think to get the OPPh3 to crash out. When you tried the hexanes/ether route, did you compound just co-crystalize and thats why it didn't work?
You might search the literature as well for a metal that chelates the OPPh3 and makes it filterable. I can't remember what metal/ligand it is off hand or the reference for that matter, but worth a search to find it while you are screening TLC conditions.