Research On Optical Mirror Testing Using Phase Retrieval With Sub-aperture Stitching Method

The modern technology in the manufacture and measurement of larger aperture aspheric optics encounters serious challenge. The large aperture of a measured aspheric optic makes the on-line measurement more important. Since phase retrieval with sub-aperture stitching method not only has advantages of simplicity, vibration insensitivity, broad testing range and capability of quantitative calculation but also promises extending the measure range of phase retrieval, it is hopeful as a technology to on-line test large aspheric optics. This thesis begins with the phase retrieval theory. According to the characteristics of aspheres and sphere we make the sub-aperture stitching programs respectively which can be used to be the theory base for on-line testing of aspheric optics. The major research efforts include the following points.1. The method to extend the measure range of phase retrieval in term of amplitude is presented. One method we bring a double-CGH phase plate to the optical path. This plate can produce some vary aberrations expressed by Zernike polynomial. These aberrations can compensate the significant departure from the spherical surface, which certainly extend the measure range of phase retrieval in term of amplitude. Another method is intensity pattern stitching which is equal to using a big size CCD to capture a full intensity pattern. The influence of misalignment between the adjacent sub-apertures on stitching is analyzed. The correctness and validity of this method are verified by an experiment.2. The preparative works for the phase retrieval with stitching is studied. In theory the relationship between the parameters of this optical path, such as the refractive index, the thickness and the position of beam splitter and the position of sub-aperture, and the support information of each sub-aperture is worked out firstly. Then a method for calibrating the position error of phase retrieval with transverse translation diversity is presented. This method comes to the conclusion that the position error of the object function results in the coordinate translation in the measurement plane but keep the amplitude constant. For calibrating this error the relationship between position error and iterative error should be clarified. In this thesis, we have proved zero position error is the minimum point of iterative error function. The computer simulations used to prove the validity of this method are described. Finally we present a means to align the optical path which is divided into the alignment between point source and tested mirror and the alignment between tested mirror and CCD camera.3. The mathematical model of the stitching method is established. Then the gradient of the error metric with respect to any parameter which can be the input of the gradient search method. The programs for stitching aspheres and spheres proved by simulation are written. In the equipment of stitching aspheres we introduce a big f/# point source which must be in the measure range of CCD camera. The position of this point source in this system has been controlled by a tilted rotation and some movements which cannot avoid introducing position misalignments influencing the precision of stitching. In this stitching method the position misalignments have been calibrated at the beginning and the end of the stitching program which also can be manipulated circulative until error metric is enough small.4. The stitching methods of aspheres and spheres are used to test a spherical mirror and a paraboloid in experiment respectively. After the comparison between the stitching results and the corresponding interferometer testing result the validity of the method is proved.5. Firstly this stitching method of phase retrieval has been used in on-line measurement of aspheric optics. The configuration for this measurement is designed. This equipment only includes a point source composed by a He-Ne laser bean and some lenses to produce a spherical wave, a beam splitter and a CCD camera. To illumination the axis departure region of tested mirror the y rotation, the x movement and the z movement of point source and the z rotation of tested mirror are needed. The values of these motions have been analyzed in theory.