Research On The Key Technology Of Defect Detection Of Large Aperture And High Precision Optical Elements

Surface defect is an important index for the quality evaluation of high precision optics.Surface defect causes light scattering,resulting in diffraction fringes,thermal aberration,coating damage,energy concentration,et al.It will even degrade the imaging quality and shorten the lifetime of high-precision optics.As the critical part of the lithography machine,the projection objective lens has strict requirements for the imaging quality.Any tiny defects on the surface will dramatically deteriorate the final exposure imaging performance of the objective lens.In order to realize impersonal quantitative detection and evaluation of micron scale defects on such large-diameter optics,an efficient and reliable automatic detection system of optical surface defects is needed.At present,the main method of defect detection is manual visual method,which is subjective,unreliabile and unrepeatable.This method is not suitable for bulk detections.Although a variety of defect detection methods are mentioned in literatures,most of them can not satisfy the requirements of large aperture,high efficiency,multi-lens type and high precision quantitative detection simulantiously.Microscopic scattering dark-field imaging method balances the needs of these aspects.It has the advantages of simple system structure,visuallized detection results and high efficiency.This method has become a mainstream one for automatic surface defect detection.However,the theory of defect scattering imaging is in great need of perfection.The system selection and parameter setting mostly rely on experience or geometrical optics simulation,which is not suitable for micron-scale defects nor can provide effective guidance for high-SNR image acquisition.When stitching subapertures of large-diameter elements,due to mechanical motion errors and the few characteristics of features in the overlapping area,existing stitching methods are facing problems such as low stitching efficiency,mis-stitching,position deviation et al.Facing the needs of large-aperture and high-precision optical surface defect detection,a high precision optics defect detection system is designed and built based on the microscopic scattering dark-field imaging method.The key problems of high-SNR imaging and efficient accurate subaperture stitching are solved.The specific work includes:(1)This dissertation discusses the significance of surface defect detection for large-diameter high-precision optical components,especially for lithography projection objective lens.The dissertation compares the standards of precision surface defect detection,and explains the reasons why these standards are not suitable for automatic defect detection.The research progress of defect detection methods and detection instruments used in the industry are introduced.Problems of surface defect detections are summarized.(2)This dissertation introduces the principle of microscopic scattering dark-field imaging method and system structure of precise surface defect detection.The two key problems in realizing high-precision defect detection of large-aperture optics are analyzed.One is how to obtain dark-field microscopic images with high-SNR,and the other is how to obtain full-aperture images quickly and accurately.Through establishment of surface defect scattering simulation model and analysis of noise,the method to improve SNR of the dark field microscopic imaging is clarified.Sub-aperture scanning and stitching are optimized to get the accurate full-aperture image.(3)The surface defects scattering model based on finite difference time-domain method(FDTD)is built.The simulation follows three steps: near-field electromagnetic field calculation based on FDTD,near-far field projection and far-field light intensity distribution analysis.The influences of defect shape,defect depth,defect width,incident angle of light source and incident direction of light source on far-field scattering imaging are analyzed.According to the simulation results,it is proposed that ring light source with dual angels should be used.The simulation provides a theoretical reference for scattering imaging analysis,system parameter optimization,and quantitative research on defects.(4)Research on sub-aperture stitching for defect detection of large aperture optics is discussed.Problems of the existing stitching methods and the reasons are discussed in detail.A defect sub-aperture image stitching method based on target tracking and adaptive scanning path correction is proposed in this dissertation.This method not only helps to decrease stitching errors such as stitching dislocation,but also improves the detection ability of weak defect targets.It reduces the impact of motion error on the detection results.(5)The defect detection system based on microscopic scattering dark-field imaging method is introduced.The defect scattering analysis experiment was carried out.Comparison between the real data of the detector and the simulation results verified the FDTD model.The stitching method based on target tracking and adaptive scanning path correction was carried out to obtain the full aperture image and repair the misalignment of the digs and scratches.Comparion between this stitching method and other methods showed that it was effective in reducing mismatches and decreasing location deviations.This defect detection system was used to test and evaluate a large aperture lithography lens,and compared with manual visual method.(6)The work of this dissertation is summarizedm,and the look farward of this research work is prospected.