Theoretical and experimental analysis of material removal and surface generation in novel fixed abrasive lapping of optical surface

Abstract

Abstract Fixed abrasive machining, which overcomes the lack of determinacy and efficiency of current loose abrasive polishing or lapping processes, is an alternative technology for fabricating large-aperture aspheric surfaces with high finishing efficiency and high-quality surface finish. A novel fixed abrasive lapping (NFAL) tool, which combines computer-controlled optical surfacing and conventional fixed abrasive lapping process, is introduced in this study to produce off-axis aspheric surfaces efficiently while surface state is controlled and subsurface damage (SSD) is avoided. The removal mechanism analysis shows that fused quartz glass is primarily removed in ductile regime and the SSD depth is 14.7 μm when the tool load is 25 N. The material removal volume is linear with the machining time under the ductile removal mode. The material removal and surface generation models are developed on the basis of the calculation of the spatial distribution of abrasive particles, the pad–particle–workpiece interactions, the single-particle abrasion mechanism, and the linearly cumulative removal of surface generation in the NFAL process. The models are verified through a series of spot and surface lapping experiments. Results show that the theoretical model can be successfully used to predict and optimize the NFAL process. The spot lapping experiments indicate that the material removal volume is linear with the rotation speed, and the maximum depth is limited to the specific value that depends on the stiffness of lapping tool. The surface roughness Ra and Rz values of the measured and simulated surface data decrease with the increase in feed rate and increase with the increase in raster spacing. The power spectral density analysis indicates that high feed rate can distinctly improve the high-frequency errors, and the selection of the raster spacing can principally affect the low-frequency and middle-spatial frequency errors.