Ionenbundelmetrologie voor gestructureerde materialen

A wide range of scientific and technological fields now rely on the ability to precisely engineer micro- and nanoscale structures through multidimensional downscaling. Among the various domains, microelectronics exhibits the most advanced and aggressive implementation of 3D downscaling. In this regard, metrology plays a pivotal role in advancing technology. Yet, the persistent miniaturization of devices featuring intricate 3D structures poses substantial challenges. In the realm of metrology, the extremely small volumes to be probed within these structures often result in inadequate sensitivity. Additionally, while nanoscale-resolution techniques enable the examination of individual structures, their limited throughput impedes statistically significant assessments across large, structured areas. To address these challenges, complementary ensemble methods have been introduced. These methods leverage large-area techniques to derive ensemble-averaged insights into the structures. By capitalizing on the high uniformity of semiconductor processing, a large collection of identical structures is analyzed simultaneously leading to a high sensitivity and statistically meaningful data. Presently, existing methods lack the capability to quantitatively derive the 3D elemental profile of nanoscale structures while preserving high sensitivity. This thesis addresses a significant challenge in metrology: achieving quantitative, precise, accurate, and traceable 3D analysis of the composition in cases of non-conformal and/or ultra-low coverage or concentration on nanopatterned surfaces with feature sizes in the 10-50 nm regime. We present ensemble Rutherford backscattering spectrometry (RBS) as a viable solution for this challenge by advancing the scientific understanding of the application of RBS to regular patterns of structures much smaller than the probing ion beam. We demonstrate 3D site-specific quantification of non-conformal and ultra-low coverage on nanopatterned surfaces using RBS. Leveraging the energy loss phenomena of the probing ions yields a localized, site-specific quantification of the composition without compromising sensitivity. By systematically exploring the interplay between experimental geometry, pattern properties, and backscattering spectra, we have developed ensemble methods based on this concept. This method enables accurate 3D quantification of elemental distributions within patterned surfaces with a sensitivity down to 1013 atoms/cm2. Our studies further explore the sensitivity, uncertainty, and dimensional limits of RBS for 3D quantification of nanopatterned surfaces. The ensemble method approaches the sensitivity of RBS as applied to planar samples while probing structures orders of magnitude smaller than the ion beam. Compared to transmission electron microscopy - energy dispersive X-ray spectroscopy (TEM-EDX), which is the industrial standard for compositional analyses on nanostructures, ensemble RBS demonstrates higher sensitivity due to its ability to probe many structures simultaneously. Even though ensemble RBS introduces significant additional sources of uncertainty compared to its application on planar samples, our investigations demonstrate that these uncertainties can be effectively managed to ensure reliable results. We also identify the dimensional limits of the methodology, which can typically not analyse structures below 15-25 nm without the use of high-resolution detectors. To overcome the challenges regarding the data analysis for 3D compositional profiling of nanopatterns, we employ artificial neural networks to speed up the quantification process and allow for a more complete analysis of the structures. This approach significantly reduces the analysis time and enables self-consistent analysis of multiple spectra, facilitating the deconvolution of complex backscattering signals from different sites within nanostructures. Overall, this thesis describes the realisation of 3D site-specific quantification in non-conformal and ultra-low coverage scenarios on nanopatterned surfaces based on RBS, paving the way for advanced ion beam metrology of patterned materials.