Accurate infrared interferometric thickness measurement of Silicon carbide epitaxial layers using dynamic refractive index correction

Read full text
Authors: Shengchen Yang, Zhen Wang, Xiangsen Han, Shuohan Ge, Dongliang Zhang (752677), Yuxuan Zhang
Cited by: 1
References: 46

Abstract

Silicon carbide (SiC) epitaxial layers are critical components in wide-bandgap semiconductor devices, where precise thickness measurement is essential for performance and reliability. Conventional infrared interferometry typically assumes a constant refractive index, leading to systematic errors due to material dispersion and carrier effects. This study proposes a physics-consistent framework that integrates Sellmeier dispersion theory with the Drude model to establish a dynamic refractive index for SiC. An analytical thickness expression is derived, reducing relative error by 2–5% and decreasing systematic deviation from 3–5% to 1–2% compared with fixed-index methods. To improve spectral peak extraction, an enhanced FFT-based approach incorporating Hanning windowing, zero-padding, and multi-band partitioning is developed. Measurements at incident angles of 10° and 15° yield consistent thickness estimates, with weighted averaging further improving stability. For multi-beam interference, a compensation strategy based on Fabry–Pérot theory is introduced and validated using silicon reference samples, achieving a thickness of 3.981 ± 0.013 μm and reducing systematic error by 10–15%. Comparative results demonstrate that the proposed method achieves lower systematic error (1.1–1.8%) and improved multi-angle consistency, indicating its effectiveness for high-precision thickness metrology and its applicability to other semiconductor and thin-film materials.

Fields of Study & Subjects

Advanced Fiber Optic Sensors Advanced Surface Polishing Techniques Optical measurement and interference techniques

Publication details

Journal / source
Nondestructive Testing And Evaluation
Publisher
Informa UK Limited
Pages
1–26
Publication date
30 March 2026
Language
EN