> Plasma ionization, can it be cheap and scalable?

It doesn't need to be awfully cheap - those elements (except iron) are pretty valuable per kg, so they can pay for a pretty expensive process while still being cheaper than getting new stuff out of the ground.