Important considerations when using the Shack-Hartmann method for testing highly aspheric optics

Abstract

The Shack-Hartmann (S-H) method is a good candidate for general aspheric metrology because the lenslet array can be designed to accommodate the dynamic range associated with wildly aspheric wavefronts. However, when the S-H method is used in this fashion several issues must be taken into consideration. First, while the sensitivity and dynamic range of the instrument can be increased by allowing the spots to shift several lenslet sub-apertures, real lenslets are not thin lenses with zero aperture so the spots will not shift in exact proportion to the average phase gradient across the lenslet as is commonly expected. Second, if the wavefront is sufficiently aspheric, any relay optics will induce additional aberrations, which can be accounted for with proper calibration and reverse raytracing. Another limitation of the S-H method is that spots cannot overlap or cross. While this is a limitation on the divergence of the phase distribution or wavefront curvature the problem can be avoided if we guarantee that the beam has no caustic between the lenslet array and detector. Finally, the single biggest problem in aspheric metrology is losing the light or vignetting. One general way to address this problem is to image the part onto the lenslet array with a large numerical aperture. In this way, rays leaving the part can have some range of angles that are guaranteed to make it through the system. This presentation will discuss these issues and methods for overcoming them. Experimental results will also be presented to demonstrate the effects.