Objective Fused silica glass has been widely used in high power laser systems and advanced optoelectronic devices, because of its good optical, thermal, and mechanical properties. The traditional optical fabrication methods are easy to cause scratches, cracks, and defects on the surface of fused silica, making the fused silica optical elements become the bottleneck restricting the development of high power laser systems, optical systems, and advanced optoelectronic devices. It is difficult to process small and complex faceted fused silica optical elements due to its high cost and time consumption. Atmospheric pressure plasma polishing (APPP), as an alternative non-contact processing method, has the advantages of non-destruction, high precision, high efficiency, low cost, and improved flexibility. The essence of APPP is a chemical reaction to generate volatile gas and achieve non-destructive removal. At present, it has been shown that it has the potential to solve the problem of processing fused quartz optical elements. However, in the actual processing process, the processing results always present the phenomenon of low convergence rate due to the different removal amount in unit time caused by the different dwell time at each processing point in the process. In this study, a variable removal function is proposed to solve the problem of nonlinear removals. The specific method is to change the single removal function in the traditional APPP process into a variable removal function that varies with the relative residence time. We hope that the nonlinear effect can be eliminated by using the variable removal function, and the actual convergence rate and machining accuracy are improved.
Methods In this paper, a 120 mm×65 mm×10 mm fused silica optical element is used as the research object. Firstly, the processing parameters such as flow rate and power are optimized to obtain a stable processing removal function. Second, several groups of removal functions are obtained on the fused silica element at different speeds, which confirms the existence of a nonlinear effect. The relationship between the nonlinear effect and the Arrhenius formula is verified by the temperature field distribution of the element surface measured by the thermal imager and the simulation analysis. Third, it is necessary to explore whether the nonlinear changes of processing residue and removal amount with speed are related or not. In the optical element surface deposition layer, the residue increases from left to right, and a removal function experiment in this deposited layer is performed. Finally, the amount of removal under different residual thicknesses is extracted by a profilometer. The correlation between the residuals and the nonlinear effect is determined according to the experimental results. After that, the coefficients of several groups of removal functions at different speeds are extracted and fitted, and the variable removal function is written based on the fitting function. The variable removal function is used for the actual processing, and the processing effect is compared with that by the original single removal function to verify whether the processing effect is improved or not.
Results and Discussions Firstly, the processing parameters are optimized, and each gas flow rate is balanced to make the material removal amount reach a stable state. At this time, the optimal parameters are the He flow rate of 2.5 L /min, the CF4 flow rate of 70 mL/min, the O2 flow rate of 50 mL/min, and the power of 170 W (Fig. 10). The experimental results confirm that the removal amount presents a nonlinear trend with time at each feeding speed (Fig. 5). After extracting and fitting the coefficients of each removal function, it is found that both the fitting function and the Arrhenius function are inverted exponential functions,and it is confirmed that the removal function is positively proportional to the Arrhenius function (Fig. 9). The influence of residues during processing on subsequent removals is compared with the removal amount under each deposition amount extracted by a profilometer. It is found that there is little difference between the removal amounts, which proves that the influence of sediment on the nonlinear effect of removal amounts is not obvious (Fig. 8). The fitting variable removal function is used for actual processing and compared with the traditional single removal function processing experiment. Compared with the processing speed distribution before processing, it is found that the speed distribution of the variable removal function is more average and clustered. After actual processing, the PV convergence rate increases from 21.4% to 65.6%, the RMS convergence rate increases from 24.1% to 72.7%, and the final processing result reaches 110λ, which perfectly completes the ultra-smooth processing of optical components (Fig.12).
Conclusions In this study, it is found that the convergence rate of fused silica processed by atmospheric pressure plasma polishing is not high due to the nonlinear effect of removal functions with different dwell times. The reasons for the nonlinear effect are analyzed, and it is verified by experiments that there is little relationship between the residue and the nonlinear effect in the processing. A new machining method is proposed to change the traditional single removal function into a variable removal function. The processing algorithm of the variable removal function is applied to the actual processing. The control experiment discloses that the PV convergence rate changes from 21.4% to 65.6%, and the RMS convergence rate changes from 24.1% to 76.7%. The PV value of the final surface shape of an optical element reaches 110λ, indicating the realization of ultra-precision machining of optical elements.
