Local-to-Global model for the surface topography prediction of laser polishing

Laser polishing has emerged as a promising method for surface polishing and structural recovery due to its significant advantages, particularly in the manufacturing of optical components with stringent surface quality requirements. However, the lack of Local-to-Global Model for calculating surface topography has hindered the intelligent integration and widespread commercial adoption of laser polishing technology. In this research, a prediction model is proposed, which enables highly convenient calculation of the global topography of laser polishing based on the energy distribution of a single laser spot. The model comprehensively takes into account plastic deformation, surface tension, viscous dissipation, and thermal capillary effects and the viability is verified through pulse laser polishing experiments on single crystal silicon and corresponding finite element simulations. The root mean square error of predicted profiles was between 0.32 – 0.61 nm in single line polishing experiments, which accounted for less than 15 % of the peak-to-valley value, and the average coefficient of determination was greater than 0.8. The median prediction accuracy for surface roughness across various frequency bands reach 80 % in multi-line superimposed polishing experiments. Dimensionless analysis led to the development of a phase diagram illustrating the relationship between typical surface topography and laser fluence, identifying the fluence range within which laser polishing technology can achieve optimal surface quality. Within this optimal laser fluence, laser polishing technology can substantially reduce overall roughness by up to 71.9 %, with a reduction of up to 91.5 % in the roughness of the primary frequency band. These findings offer crucial theoretical support and practical guidance for the intelligent control and industrial application of laser polishing in high-precision optical manufacturing.