When using PMMA for e-*beam applications, you can try working with two layer
system, having some lower MW PMMA at the bottom and high MW PMMA at the top.
When you say, that your resolution is just about 500 nm, that should work.
Example: 200 nm PMMA 250K and 300 nm PMMA 950K.
Other thing you can do is using DMFA (di-methyl-form-amid) instead of
acetone. This chemical doesn't dissolve the resist, but kind of peels it off
the substrate including the metal. I use it sucessfully to lift off just
20-20 nm small features and it works perfectly fine.

Send me an email if you need some more help.

greetings

Josef Kouba
josef.kouba@bessy.de