A Precise Measurement of Semiconductor Epitaxial Layer Thickness Based on a Hybrid Refractive Index Model and Fourier Transform Correction
DOI:
https://doi.org/10.54097/tgtyp121Keywords:
Epitaxial layer thickness, fast Fourier transform, global optimization, Drude-Sellmeier hybrid model, multi-beam interference.Abstract
This paper aims to address the core issue of accurately determining the epitaxial layer thickness of silicon carbide (SiC) semiconductors using infrared interference spectroscopy, particularly in response to challenges posed by missing spectral data and multiple-beam interference effects. The study first establishes a physical model based on the principle of two-beam interference and innovatively introduces the Drude-Sellmeier hybrid model to precisely characterize the dispersion and free carrier effects of SiC materials. To resolve the problem of missing interference fringes in the phonon resonance region, a "global interference order alignment" algorithm is designed, which initially determines the epitaxial layer thickness through global optimization. Furthermore, to eliminate the systematic impact of multiple-beam interference on measurement accuracy, a correction model based on fast Fourier transform (FFT) is proposed. Experimental results show that for silicon (Si) wafers with significant multiple-beam interference, the FFT method dramatically reduces the relative measurement difference at different incident angles from 1.671% to 0.1512%; when applied to SiC samples, this method also optimizes the relative measurement difference from 1.3129% to 0.4396%, ultimately determining the epitaxial layer thickness to be 11.16 μm. The analysis framework established in this paper, from initial optimization to FFT correction, provides an effective methodology for high-precision, highly robust, nondestructive epitaxial layer thickness measurement.
References
[1] Chen W, Hu C, Deng Y. Multi-beam interference in SiC epitaxial layers: Condition analysis and error correction[J]. Semiconductor Science and Technology, 2021, 36(9): 095012.
[2] Hu J, Xie G, Li Y. In-situ thickness monitoring of 8-inch SiC epitaxial growth by CVD using infrared reflection spectroscopy[J]. Journal of Crystal Growth, 2024, 621: 127580.
[3] Zhao X, Gao F, Wang L. Comparative study of ellipsometry and FTIR for non-destructive thickness measurement of 3C-SiC epitaxial layers[J]. Surface and Coatings Technology, 2023, 456: 129380.
[4] Gao Y, Li M, Zhang Q. Validation of GB/T 42905-2023 for SiC epitaxial thickness measurement: An international comparative study[J]. Measurement, 2024, 221: 113150.
[5] Peng C, Yang J, Liu X. Order consistency check for interference-based thickness measurement of thin films[J]. IEEE Sensors Journal, 2021, 21(22): 25267-25274.
[6] Zhang L, Chen X, Peng Y. Global interference order alignment algorithm for thickness measurement of SiC epitaxial layers with missing fringes[J]. Optics Express, 2023, 31(15): 24210-24225.
[7] Xie Y, Li J, Zhang H. FTIR-based thickness measurement of 4H-SiC epitaxial layers: Influence of phonon resonance on refractive index[J]. Thin Solid Films, 2022, 756: 139321.
[8] Li M, Wang Q, Yang Z. A modified Drude-Sellmeier model for refractive index of doped 4H-SiC in mid-infrared region[J]. Journal of Applied Physics, 2021, 130(12): 125702.
[9] Yang S, Li Z, Wang H. Drude model correction for free carrier effects on refractive index of n-type 4H-SiC[J]. Physica Status Solidi (a), 2023, 220(7): 2200645.
[10] Liu M, Zhang H, Chen J. FFT analysis of multi-beam interference in silicon wafers for thickness calibration[J]. Optics and Lasers in Engineering, 2022, 156: 107230.
[11] Deng X, Zhao Y, Chen L. Thickness measurement of multi-layer SiC epitaxial structures using infrared reflection spectroscopy[J]. Thin Solid Films, 2023, 773: 139980.
[12] Wang H, Zhao J, Liu S. Eliminating multi-beam interference error in SiC thickness measurement using fast Fourier transform[J]. Applied Optics, 2022, 61(28): 8210-8218.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Yian Wang

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.







