Study On Theory And Technology In Ion Beam Figuring For Optical Surfaces

With the ever-increasing demands on performances, optics used in modern optical systems such as weapons, telescopes, laser fusion systems and extreme ultraviolet lithography (EUVL) systems demand higher accuracy and more quantity. The present widely used computer controlled optical surfacing (CCOS) technique, although gains more efficiency than conventioanal polishing methods, can not satisfy the critical demands due to its lower accuracy and lower convergence ratio. Moreover, it often causes high-middle frequency error and edge effects on optical surface. Ion beam figuring (IBF), which removes surface material by physical sputtering, hold very stable beam removal function (BRF) due to its unique manner to remove material. It is insensitive to machining distance, surface curvature, environment vibration and optics support. Moreover, its BRF is in Gaussian shape and is free of edge effect. Consequently, IBF process can obtain high accuracy and high convergence ratio than other deterministic process, which makes it an effective and applicable technique to produce high accuracy optical surfaces, especially aspheric surfaces.Although IBF process has been successfully applied since the early 90s, there is little work on it at home due to its relative complexity and its high cost. Because it is a key technology to produce high precision optical surfaces used in both military and commerce, we should develop it. This thesis is dedicated to the theory and technology of IBF process, including removal function, dwell time, error correcting ability, error sources, processing technique and plentiful experiments. The major research efforts include the following points.1. The BRF in IBF process is modeled and studied. First, the theoretical model of BRF is founded using the Sigmund theory. Then the characteristics of BRF are investigated according to the model. The methods to estimate BRF by experiments are discussed and a new method named line scan method is proposed. Using line scan method the stability of BRF can be investigated. Finally, some experiments are performed on KDIFS-500 to test the shape, stability and conformability of BRF.2. In order to determine dwell time, which is the key input variable for successful figuring processes, two approaches are investigated with emphases on edge effect and parameter optimization. First, the CEH model is introduced, which can completely eliminate edge effect. The TSVD regularization algorithm is also introduced in order to solve the CEH model, and to determine a reasonable regularization parameter in the TSVD algorithm. A new curve named figuring prediction curve is proposed and applied, which is defined as process time vs. residual accuracy. Lastly, the iterative method to determine dwell time is also investigated due to its less computing time. Some improvements are proposed to reduce edge effect and to determine a reasonable dwell time, which makes the iterative method more practical.3. To evaluate the correcting ability of a figuring process to different spatial frequency figure error, a quantitative criterion named material removal availability (MRA) is proposed, which is defined as the ratio of the volume of desired material removal to that of the real material removal. First, the correcting ability of a Gaussian BRF is investigated and the investigation indicates that the correcting ability of an IBF process is determined by the diameter of the ion beam, and the quantitative relationship is given. Then, the correcting ability of an arbitrary BRF is investigated. Furthermore, the correcting ability of circular symmetrical BRF and the correcting ability of one dimensional figuring process are discussed. Lastly, three sine figures are etched by IBF processes and the figuring results agree with the theoretical predictions.4. The major error sources, the positioning error and the BRF's error, are discussed based on the two-dimensional convolution model. The investigation shows that the residual error resulted from positioning error is the dot product of the figure error gradient vector and the positioning error vector, and the residual error resulted from a constant bias error of BRF is proportional to the material removal.5. Several improvements are made on processing technique. A positioning method is proposed and realized in order to reduce positioning error. The velocity process mode is realized in order to reduce processing time. And a novel spiral scan path named uniform-area-growth-ratio spiral path is proposed and realized to optimize processing path. These improvements are of great benefits to IBF process with higher convergence ratio and less residual error.6. Finally, IBF experiments are performed on a variety of optical surfaces including flats, sphere and aspheres, on materials including Zerodur glass and SiC, and on shapes including circle and ellipse. The final residual errors after IBF process are all less than 10nm RMS (the best value is 3.1 nm RMS). The successful figuring results prove the validity and advantages of the proposed algorithms and the proposed process improvements.