I am looking into techniques to coat a silicon wafer with copper with a variable topology thickness (e.g. z=u*(sin(f*x)+cos(f*y))). My current idea involves sputter coating using an arduino controlled xy-table in high vacuum and a laminar flow nozzle to get a pinpoint stream of atoms.

Things which I have not fully considered yet is the geometry of the flow nozzle (cone, cylinder, something else), coating the flow nozzle with a material which doesn't let copper atoms adhere (like a material which has very different crystal lattice constants?) and possibly using further magnetic or electrical fields to guide the copper atoms towards the centre of the nozzle.