Optical polishing process: Analysis and optimization using response surface methodology (RSM) for large diameter fused silica flat substrates

The skill dependent optical polishing process remains elusive due to unaccounted parameters and uncertainties. Deterministic relations for surface quality of polished optics have been developed in this paper by analysis and optimization of three significant input parameters: abrasive concentration, relative velocity and abrasive size. Average roughness value (Ra), measured on a surface profilometer: Talysurf PGI-120, indicated the response. MALVERN ZETASIZER was used to measure average size of abrasives using Dynamic Light Scattering principle. The parameter values have been computed in standard units for general relevance across different setups, a feature absent in earlier studies. Flat Fused Silica glass samples of 100 mm diameter were polished for two hours on a conventional single arm spindle polisher with a polyurethane pad using Ceria slurry. Central Composite Design (CCD) of RSM was used for fitting a second order model. Kinematic analysis established the polishing spindle and sample rotation speeds to be equal for uniform velocity distribution. Relative velocity was the most significant input, followed by concentration. Surface and contour plots resulted in a domain of peak performance. Optimum combination (velocity 1054.8 mm/s, abrasive size 8.5 nm, concentration 2.45%) identified using desirability approach based on the second order regression model resulted in Ra value as low as 5.7 nm, well within permissible error limits from the predicted value.