| Laser is characterized by its long coherence length and has found wide applications in the field of interferometry. However the coherence speckles generated by the scattered or unwanted reflected light affect the accuracy of the measurement, even make the measurement failed. Besides phase-shifting interferometry using laser suffers from phase-ambiguity problem. The coherence length of a broadband light source, especially white-light source, is short because of its large spectral bandwidth, and the high-contrast fringes are obtained only when two path lengths of the interferometer are closely matched in length, therefore the spurious fringes are not easy to formed. At the same time, the phase-ambiguity problem is avoided because the zero-order fringe is obvious in white-light interferometry. Furthermore the white-light fringes can provide abundance color information. As mentioned above, the interferometry of broadband light source becomes more and more significant. The main work is based on the theory of white-light interferometry, which is described as follows:First, white-light interferometry is used to measure the retardation of birefringence crystal. A method of measuring retardation of zero-order wave plate is presented based on the hue of white-light interference color. The relations between retardations and its corresponding hue values are calibrated by using a Soleil-Babinet compensator. Besides, a method to calibrate the Soleil-Babinet compensator is presented. Therefore the retardation of a test wave plate can be obtained by the hue value of its corresponding white-light interference color. The retardation of the zero-order wave plate is measured in the experiment, whose result (564.9nm) coincides with the one obtained by spectroscopic method (565.2nm). It can also be used to measure the stress of optical glass.A method using white-light Michelson interferometer for measuring the retardation of wave plates is also presented. The linear polarized white-light splits into o-beam and e-beam by the test wave plate which introduces retardation between the two beams, and then they are divided by a beam splitter and reflected by two plane mirrors of the Michelson interferometer respectively. Finally three white-light interference packets are formed. For a multiple-order wave plate, the interference packets will be separated absolutely, and according to the optical path between the center packet and one of the side packets, the retardation of multiple-order wave plates can be obtained. The retardation of the multiple-order wave plate is measured in the experiment, whose result (2990.6nm) coincides with the one obtained by spectroscopic method (2992.8nm). The retardation of a test lower-order wave plate or the stress of a optical glass can be calculated by using the phase change after inserting it behind the multiple-order wave plate.Second, the white-light interferometry is used in the measurement of surface profile. The Linnik and Mirau type white-light interference vertical scanning systems are separately built on the 6JA and metallurgical microscopes. The white-light interference signals are processed by the spatial frequency domain algorithm while the test sample is vertical scanned by PZT. The step height of a step surface and the trench of a CGH (Computer Generated Hologram) are measured whose results coincide with the ones from ZYGO NewView 7200 profiler and Ambios XP2 stylus profilometer.At last, in order to test the homogeneity of optical crystal with parallel surfaces, a low coherence light source is used on the Twyman-Green interferometer. Both the front and the rear surface of the tested sample will be placed close to the zero-OPD position by adjusting the optical path of the reference arm. Then the lights reflected from the two surfaces will not interfere with the reference light simultaneously. So the lights reflected either from the front surface or from the rear surface can be distinguished. The interferograms are processed with the virtual grating phase-shifting Moire fringe method. The measurement result indicates that the accuracy is within 2.6×10-6, the peak-to-valley value isΔnpv=6.06×10-6, and the rms isΔnrms=8.96×10-7. |
PDF Full Text Request