The surface roughness of silicon wafers is a critical factor affecting semiconductor device performance. Light scattering methods based on the bidirectional reflectance distribution function (BRDF) are widely used for roughness analysis. However, conventional angle-resolved detection equipment required to obtain BRDF is structurally complex, costly, and time-consuming. Numerical simulation provides an effective solution to obtain BRDF from silicon wafer surfaces. In this study, field emission scanning electron microscopy (FE-SEM), atomic force microscopy (AFM), and white light interferometry (WLI) were used to characterize polished silicon wafer surfaces. The presence of residual polishing particles on the wafer surface was confirmed. These particles create depressions with directional features, exhibiting anisotropic, non-Gaussian, and fractal characteristics. A non-Gaussian fractal surface model was subsequently established using the spectral representation method (SRM). Finally, BRDF of these surfaces was obtained using the finite-difference time-domain (FDTD) method and vector perturbation theory. This study provides an effective way to analyze the scattering of silicon wafer surfaces.
Li et al. (Thu,) studied this question.