Methods for automatically obtaining the sputtering-induced roughness upon depth profiling of polycrystalline films

Sputter depth profiling has been widely used for analyzing the in-depth compositional distribution in thin films. However, due to the unavoidable distortions in sputter depth profiling, the quantitative analysis of measured depth profiling data requires a modeling approach. By estimating parameters such as the original film structure, roughness, elemental sputtering rate, and mixing length, we can fit a model to the experimental data and correct for these distortions. This allows us to obtain a more accurate representation of the original film structure. In this paper, two methods—the simulated annealing algorithm and the golden section search algorithm—are proposed for obtaining automatically and accurately the sputtering-induced roughness that often occurs upon depth profiling of polycrystalline films. When comparing these two methods, the golden section search algorithm demonstrates superior efficiency and precision under low-noise conditions, thereby reducing computational time and error. While, under high-noise conditions, the simulated annealing algorithm achieves greater accuracy at the cost of significantly higher computational time. As an example, the measured Auger electron spectroscopy depth profiling data of 4 × Si(15 nm)/Al(15 nm)/Si(100) multilayers are used for obtaining automatically the sputtering-induced roughness by these two methods. It is suggested that the golden section method offers an effective balance between computational time and accuracy.