Accurate determination of thin film thickness by X‐ray reflectivity (XRR) can be achieved using several data analysis approaches; however, their relative accuracy, computational effort, and practical applicability require systematic comparison. In this work, the thickness of the Fe 2 O 3 /Fe system on an Al 2 O 3 substrate is determined and compared using four XRR‐based methods: peak separation, fast Fourier transform (FFT) analysis, Bragg equation modified by the total reflection condition method, and curve fitting. The thickness values obtained by all four methods are consistent within 1.34%, demonstrating their overall reliability. The peak separation method is the most efficient approach, allowing rapid thickness estimation using a limited number of interference peaks. When peaks at larger 2 Θ angles are used, the resulting thickness differs by only 1.34% from that obtained by curve fitting. The FFT‐based method yields a thickness value very close to the curve fitting result (67.52 nm versus 67.0 nm) but requires extensive data preprocessing, including identification of the oscillatory region, signal smoothing and normalization, DC subtraction, conversion to a uniform angular step, windowing, and zero padding. Despite this complexity, the FFT method is advantageous because it enables separation of frequency components associated with individual layers. The Bragg equation modified by the total reflection condition method shows a deviation of approximately 0.8% from curve fitting when both maxima and minima in the oscillatory region are considered. This method involves linear fitting of extrema positions and allows determination of the critical angle and film thickness. Curve fitting is the most time‐consuming method but provides the most comprehensive information, enabling simultaneous determination of film thickness, density, roughness, and detection of additional layers formed, for example, by oxidation. The first three methods allow one to determine the “effective” thickness of the entire layer, while the approximation method allows one to separate the contribution of the metal and the oxide formed on the surface of the film. Using thickness values obtained from simpler methods as initial parameters significantly reduces the complexity of curve fitting.
Lishchenko et al. (Thu,) studied this question.