Spectroscopic ellipsometry is a non-destructive and rapid evaluation method for multilayer semiconductor thin films. This technique measures the change in polarization state of light reflected from the sample surface across multiple wavelengths. It offers advantages over traditional methods, such as the ability to evaluate the complex dielectric function, film thickness, and surface roughness of multilayer films. The technique has been applied to evaluate amorphous silicon (a-Si) thin films, with studies by Collins showing its effectiveness. The principle of spectroscopic ellipsometry involves measuring the tangent of the psi angle (tanψ) and the cosine of the delta angle (cosΔ) of the reflected light. These parameters depend on the complex dielectric function and thickness of each layer in a multilayer structure. For inhomogeneous films, the Bruggeman approximation is used to estimate the effective dielectric function. The data is then fitted using the least squares method, with the Marquardt algorithm improving the accuracy of the analysis. In evaluating a-Si films, the complex dielectric function and film thickness were determined by analyzing the changes in the optical energy. The results showed that thinner films have lower dielectric functions at low energy. The analysis of different models revealed that the third model, which included gaps in the a-Si layers, provided the best fit to the measured data. The results indicate that the structure of a-Si films changes as the film grows, and spectroscopic ellipsometry can non-destructively evaluate these changes, including surface roughness and natural oxide layers. Spectroscopic ellipsometry is increasingly being used not only in research but also in quality control in manufacturing. Recent developments include in-situ measurement of a-Si film growth and infrared spectroscopic ellipsometry. With the adoption of new regression algorithms and software, the accuracy and efficiency of data analysis have significantly improved. This technique is highly effective for analyzing multilayer structures in amorphous materials and III-V compound semiconductors, and is expected to become widely adopted in the future.Spectroscopic ellipsometry is a non-destructive and rapid evaluation method for multilayer semiconductor thin films. This technique measures the change in polarization state of light reflected from the sample surface across multiple wavelengths. It offers advantages over traditional methods, such as the ability to evaluate the complex dielectric function, film thickness, and surface roughness of multilayer films. The technique has been applied to evaluate amorphous silicon (a-Si) thin films, with studies by Collins showing its effectiveness. The principle of spectroscopic ellipsometry involves measuring the tangent of the psi angle (tanψ) and the cosine of the delta angle (cosΔ) of the reflected light. These parameters depend on the complex dielectric function and thickness of each layer in a multilayer structure. For inhomogeneous films, the Bruggeman approximation is used to estimate the effective dielectric function. The data is then fitted using the least squares method, with the Marquardt algorithm improving the accuracy of the analysis. In evaluating a-Si films, the complex dielectric function and film thickness were determined by analyzing the changes in the optical energy. The results showed that thinner films have lower dielectric functions at low energy. The analysis of different models revealed that the third model, which included gaps in the a-Si layers, provided the best fit to the measured data. The results indicate that the structure of a-Si films changes as the film grows, and spectroscopic ellipsometry can non-destructively evaluate these changes, including surface roughness and natural oxide layers. Spectroscopic ellipsometry is increasingly being used not only in research but also in quality control in manufacturing. Recent developments include in-situ measurement of a-Si film growth and infrared spectroscopic ellipsometry. With the adoption of new regression algorithms and software, the accuracy and efficiency of data analysis have significantly improved. This technique is highly effective for analyzing multilayer structures in amorphous materials and III-V compound semiconductors, and is expected to become widely adopted in the future.