Thickness measurement is a fundamental metrology technique in the field of precision engineering and plays a critical role in various industries, including semiconductor, display, and film manufacturing. In particular, the thickness of materials such as silicon wafers is directly related to process quality, and thus non-contact, real-time measurement techniques are widely employed in industrial environments. Spectral interferometry is well recognized as a representative method that meets these requirements. However, in order to practically utilize measurement results obtained from such instruments, it is essential not only to acquire measurement values but also to quantitatively evaluate the reliability of those values. From an industrial perspective, this aspect is as important as, or even more important than, proposing a new measurement technique.
In this study, the measurement uncertainty associated with optical thickness measurements of silicon wafers was systematically analyzed using a spectral interferometer equipped with a high-speed spectrometer (t-Nova-1550, Meter-Lab Inc.). The uncertainty evaluation was conducted in accordance with the internationally recognized Guide to the Expression of Uncertainty in Measurement (GUM). In particular, three dominant sources of uncertainty were identified and analyzed: (1) analysis error arising from the discrete Fourier transform algorithm, (2) measurement repeatability obtained from repeated measurements at the same location, and (3) wavelength uncertainty of the spectrometer.
Each uncertainty component—including algorithm-induced errors during optical thickness extraction from interference spectra, system reproducibility, and wavelength-axis uncertainty originating from instrument specifications—was individually evaluated. These components were then combined to determine the combined standard uncertainty. The results of this study provide objective verification of the instrument’s measurement performance and are expected to serve as fundamental reference data for future research and development, including algorithm refinement and spectrometer performance enhancement.
