Experimental investigation of mid-spatial frequency surface textures on fused silica after computer numerical control bonnet polishing

Automated bonnet polishing is achieved using computer numerical control (CNC) technology. However, owing to specific toolpaths, CNC bonnet polishing is generally accompanied by the mid-spatial frequency (MSF) of surface textures on polished surfaces that could produce surface ripples or waviness and degrade image quality. In this study, the MSF surface textures on fused silica are investigated by developing a CNC bonnet polishing technique using a cerium oxide-filled polyurethane pad (LP66)—a cellular polyurethane material designed to handle high flatness and surface finishing requirements for optical glass materials. To minimize the MSF, optimal combinations of various polishing parameters, including tool offset, head speed, track spacing, and surface feed rate, are studied. Experimental results demonstrate that the head speed and feed rate significantly affect the surface texture during bonnet polishing. Although the tool offset does not cause surface textures, the material removal rate is affected. A series of optimization experiments is conducted, consequently leading to the effective removal of irregular surface ripples and a reduction of MSF errors. By optimizing the polishing parameters, extremely accurate surface quality is achieved, along with a root mean square error of 1.6 nm. These results demonstrate the potential applications of LP66 in CNC bonnet polishing for highly accurate freeform optical components.