A Research On Measurement Methodology Of Polarization Performance

Nowadays, integrated circuit(IC) has become the cornerstone in every high-tech field, and the increase of the IC’s integration degree never stops under the Moore’s law. Optical projection lithography is the most nucleus technology in the production of very-large scale integration(VLSI). With the precision of lithography growing smaller than 45 nm, immersion optical exposure lithography is currently prevailing in the production of IC. According to the Rayleigh diffraction theorem, the decrease of the exposure wavelength and the increase of the numerical aperture(NA) are required to guarantee the promotion of the lithographical resolution. Ar F immersion lithography can achieve numerical aperture greater than 1, and the influence it takes from polarization diffraction effect grows heavier with the increase of the NA. Polarized illumination and off-axis illumination can greatly promote the image contrast of lithography, and the corresponding lithographical system is demanding in the performance of the illumination system. Thus, only systems which can guarantee highly-precise measurements of the polarization performance can insure satisfactions from the illumination light to the polarization requirement.According to the requirements mentioned above, this paper researches on the measurement system which uses rotating wave plate to measure stokes vector for polarization index calculation. Simulation is designed and theoretical validations are carried out for the measurement of the linearly-polarized light. The main content of this paper is shown below:1. I surveyed on several polarization measurement methods and decided to use rotating quarter-wave plate to measure stokes vector. The optical transition function is derived and the least square method is adopted to solve the Stokes vector.2. For application scenarios of 193 nm Deep Ultra Violet lithographical machine, the corresponding errors generated during each error-prone part are simulated and analyzed, which include the nonuniformity of the illuminant, the angular error of the polarization prism’s axis, the zero-setting error of the rotating plate’s fast-axis, the phase delay error of the rotating plate, the influence of the motor’s rotation precision and the CCD noise.3. For the parameterized errors: the angular error of the polarization prism’s axis, the zero-setting error of the rotating plate’s fast-axis and the phase delay error of the rotating plate, a calibration method is proposed. This method uses 0° and 45° linearly-polarized light to illuminate the system for error calibration. The results of the simulation show that this method can effectively promote the measurement precision and reduce the errors.4. Experiments under 632.8nm wavelength are carried out. The measurement methodology and the calibration methodology are validated. I analyzed the influence of the errors emerged during the experiments, and present the method for improving.