Looks good to me. The only thing you won’t get is TIR from the sides of the zone plate, or rays interacting at all with the sides of the plate. You’ll need non-sequential for that.

Oh wait. The Zone plate lens will not let you define a radius per zone...it’s just the thickness of a flat plate that changes. 

There is a user-defined surface type called US_Multizone_Aspheric. This models a series of annular even aspheric zones. Each zone has a radius, conic even asphere coefficients to order 12, and a maximum radial aperture for the zone. Up to 20 concentric zones are supported. However, each zone is shifted so the surface is continuous across the zone boundaries, and you want a staircase. You may be able to modify the source code to provide this.

• Mark

@Mark.Nicholson , I haven’t made a user-defined surface before, but modification of the US_Multizone_Aspheric looks like it would be straitforward to modify. Thank you for the suggestion.

I was also thinking that I could make this surface using a grid sag. This approach wouldn’t be directly optimize-able, but it would be a good first test.

Another option would be to create the geometry in either the ‘Part Designer’ or external CAD program and import as a non-sequential object.

Any thoughts on these approaches?

I’ll pick back up on this in a few weeks, and provide an update on any progress.

Thank you,

John

I think Mark’s last answer about using a non-sequential hybrid approach is going to work out the best. For POP, you’ll need to use rays to pass through the actual surfaces or object anyway, so the easiest approach is to use NSC for what it excels at - easy creation of complicated shapes.