基于红外干涉法的SiC外延层厚度测量反演建模OA

In addition to the mixing of dual-beam and multi-beam interferences,the material dispersion effects are found in infrared interferometric measurement of SiC epitaxial layer thickness.An infrared interferometric measurement model combining physical mechanisms and data-driven optimization is proposed to construct an inversion optimization framework with the minimization of reflection spectrum residuals by establishing a forward interference model based on Cauchy dispersion.For dual-beam interference,the differential evolution(DE)algorithm is used to achieve high-precision inversion with the average thickness of 6.48 μm(10° incidence)and 5.47 μm(15° incidence),and a relative error of less than 5.17%.For multi-beam interference,a reflection threshold(R＞0.5)is derived and used as a criterion,obtaining the reduced root mean square error(RMSE)of 0.175.The results show that the proposed model helps achieve high inversion accuracy and robustness.