Thin Film Thickness Measurement And Processing Technology Of Interferogram Based On Interferometry

With the rapid development of electro-optical technology and micro-electronic devices, thin film technology has been widely applied in optical engineering, micro-electric technology, communication, astronavigation and so on. The measurement of thin film thickness which directly influences the electrical characteristics, the optical characteristics, thermal conductivity etc, has been the focus in this field. The modern interferometry measurement with high precision, high-resolution, high sensitivity, non-contact measurement, the whole test and simplicity of operation, are widely used in precision measurement. The key technology of modern interferometry is to obtain the necessary surface shape and parameters by processing interferogram with reasonable algorithms. In this paper, based on the interference principle, many in-depth researches on the thin film thickness measurement have been done, and a novel method to obtain the thin film thickness using the image processing is proposed.In this paper, the advantages and disadvantages, accuracy, scope of appliance and usable condition of many usually used measurement methods of thin film thickness which are analyzed and discussed in detail, are been summarized. By using two different interferometers (Twyman-Green interferometer and Michelson interferometer), two high-precise digital image acquisition systems involved Charge Coupled Device (CCD) receiving apparatus are separately set up. The processing algorithms of interferograms grabbed from the settled systems are mainly researched in the process of thin film thickness measurement. A series of pre-treatment algorithms of interferograms, including noise deduction, edge extraction based on Mathematical Morphology and region extension based on two dimensional fast Fourier transform iteration method are designed. After pre-treatment, the interferograms are transformed into frequency domain. Through a well-designed filter, interference pattern spectrum in level spectrum are obtained, then processed by fast inverse Fourier transform, the influence of the background interference fringes to film thickness measurements is eliminated. Through analyzing some wrapping and unwrapping algorithms, the least squares algorithm based on discrete cosine transform (DCT) is proposed to unwrap the wrapped phase of thin film samples calculated using arctangent function in this paper. According to the algorithms above, a whole set of programs are designed and calculated by MATLAB, and three-dimensional surface topography is obtained. Then the thin film thickness can be obtained by sampling analysis on the section in the surface topography including the thin film thickness information.In order to verify the feasibility and accuracy of the method, several thin film samples with different shapes are prepared. The thickness interferograms of samples obtained by the test system have been analyzed and processed by the settled algorithms, and the calculation data compared with that obtained from Zygo interferometer. The results show that algorithms in this paper are quite feasible, with high accuracy in thin film thickness measurement and less than 1% in relative deviation.