【作者】 朱硕；

【导师】 张晓辉；

【作者基本信息】 中国科学院研究生院(长春光学精密机械与物理研究所) ， 光学工程， 2014， 博士

【摘要】 随着科学技术的飞速发展，大口径光学系统在空间光学、天文光学等领域得到越来越广泛的应用。大口径平面镜常作为标准器具用于光学系统的自准检验，因此，对大口径平面镜的面形误差精度有很高的要求，同时也需要有高精度的检测手段来完成面形检测。目前商用数字干涉仪的最大口径为1m，当平面镜的口径超过1m时，无法实现全口径面形的直接检测，使大口径平面镜高精度检测成为难题，本文对几种常用的大口径平面镜面形检测方法作以总结并进行分析比较，选择瑞奇-康芒法作为主要研究内容。根据瑞奇-康芒检测原理，分析了系统波像差与被检平面镜面形误差之间的对应关系。针对瑞奇-康芒法单次检测无法得到被检平面镜实际面形情况的难点，选择采用改变主光线入射角的办法进行两角度检测。提出利用检测系统光瞳面与被检平面镜表面二者间坐标及幅值转换关系来计算平面镜的面形误差的方法。详细介绍了这种数据处理方法的计算原理，对算法的计算精度进行了分析，平面镜面形误差计算精度可达0.01(=0.6328μm)，并在最终结果能够分离出光路调整引入的误差，使计算得到的平面镜面形更为准确。通过仿真模拟实际检测光路，分别对利用系统波像差与面形误差之间影响矩阵计算面形误差的方法及利用二者间坐标及幅值转换关系计算面形误差的方法进行了研究，对比两种数据处理方法的结果，证明本文提出的利用坐标及幅值转换关系计算平面镜面形误差的方法精度更高，更有利于分析瑞奇-康芒检测数据。同时对影响瑞奇-康芒检测精度的要素进行了详细分析，利用仿真分析了瑞奇角选择范围及其精度对检测结果的影响，确定出了适于测试的瑞奇角的选择范围以及瑞奇角的测量误差允许范围。为了进一步保证瑞奇角的测量精度，提出了一种新的测量瑞奇角大小的方法，并给出了此方法的精度。对干涉仪镜头焦点到平面镜中心距离的测量精度对检测结果的影响作了详细分析。对小口径平面镜进行瑞奇-康芒验证实验，将计算得到的面形结果与干涉仪直接检测结果做对比，验证了面形数据处理方法的有效性，以及精度分析理论的正确性，为后续大口径平面镜的检测工作提供了理论依据，奠定了基础。对基于五棱镜扫描测试原理实现大口径平面镜面形检测的相关技术进行了研究，研究内容对于实现大口径光学元件面形的高精度测量具有一定的现实意义。介绍了五棱镜扫描法的检测原理及测试系统组成，着重分析五棱镜扫描测试环节的扫描精度，编制软件用于计算五棱镜存在角度制造误差以及三维运动误差时带来的角度测量的偏移量，以便在后续数据处理中进行处理。分析了扫描路线与初级像差之间的关系，确定了选择极坐标扫描采样方式进行测试更能保证在短时间内获得精度较高的面形测试数据，提高检测效率。利用分析得到的结论，搭建了1.5m平面镜瑞奇-康芒检测光路，使用的标准球面镜口径1.8m，曲率半径15m，通过测得的两角度测试波前数据分析并计算平面镜的面形误差。得到平面镜的面形结果其峰谷值PV为0.391(=0.6328μm)，均方根误差RMS值为0.0181(=0.6328μm)，并给出了平面镜的面形误差情况。最终使用瑞奇-康芒法完成了1.5m平面镜的面形检测工作，从而实现了大口径平面镜的高精度检测。更多还原

【Abstract】 With the rapid development of science and technology, large-aperture opticalsystems are more widely used in space optics, astronomy optics and other fields.Large optical flats are often used as reference surfaces in autocollimation optic tests.Thus, a high-quality surface and a suitable and highly precise test method isnecessary. Currently the largest commercial digital interferometer of1meteraperture is not available to cover the large mirror which diameter is more than1mduring tests. This paper analyzes some common method which widely used in largeaperture flat mirror test, and select Ritchey-Common test method as the maincontent.According to the theory of Ritchey-Common test, this paper analyzes therelationship between the system wavefront and flat mirror surface. To solve thedifficulties in the Ritchey-Common test that a single measurement could not obtainthe test flat mirror surface, this paper change the main optical path to derive thesurface from two different test angles. And we utilize the relationship betweensystem pupil coordinate and flat mirror coordinate to derive the surface. The detailsand accuracy of the data processing method is well presented. The accuracy canreach to0.01(=0.6328μm), a more actual flat surface error could be obtain forthe alignment error of optical path separated.Compared the methods of using the influence matrix with our method of using relationship between system pupil coordinate and flat mirror coordinate to figure outthe flat surface by simulation.The result shows that our method is more accuracy andsuitable for Ritchey-Common test. Asuitable range of the Ritchey angle is identified,and its error allowance is determined by simulation. The ratio of image size to thepupil plane is used to calculate the Ritchey angle in test, the accuracy of this methodfor calculating is high enough. The distance between the lens focus and the flatmirror center is analyzed. We test a flat mirror with a small diameter in threedifferent Ritchey angles. The results of the Ritchey-Common test and the directmeasurement with Zygo are very close; the experimental results confirm that thisRitchey-Common method is effective and accurate.And this provided analysis foractual tests.This paper also introduces the composition and principle of the scanningpentaprism test system which is very important for large aperture optical test. Themethod could be used to test large flat mirror which diameter is larger than1m.This paper focuses on analyzing scanning accuracy of the pentaprism, andprogramming software that used to calculate the manufacturing errors of thepentaprism and the angle error when moving in three-dimensional direction. Theresults are used in the subsequent data processing. We also analyze the relationshipbetween the scan lines and the primary aberrations, and select the polar coordinatescan sampling method for testing which is more accurate and efficiency.We build the Ritchey-Common test path for a1.5m flat mirror. A well spherewith the diameter of1.8m and the radius of curvature is15m was used as areference. We derive the surface from two different test angles. In the results, PV andRMS can reach0.391(=0.6328μm) and0.0181(=0.6328μm) respectively.The test of1.5m flat mirror is finished by Ritchey-Common method. Theexperimental results confirm that this Ritchey-Common test method is effective andaccurate.更多还原