Fig. 5

Setup characteristics and resolution/reflectivity trade off diagram. a The characteristics of the SXCT setup are summarized in a comprehensive table. Due to the nonlinear relationship between wavelength and photon energy, the center wavelength does not correspond to the center energy—only the spectral boundaries do. The lateral resolution is defined by the focal spot size. The axial resolution was determined using a Gaussian fit to the reconstructed depth structure. The sensitivity was determined on basis of Fig. 4c. The imaging depth span is defined by the spectrometer resolution, which was estimated from the illumination spectrum or can be calculated based on the pixelation at the highest photon energy used (in brackets). b The diagram illustrates the trade-off between axial resolution (blue) and reflectivity (orange) as a function of the angle of incidence \(\theta\). The theoretical resolution was calculated assuming a rectangular input spectrum. Typical commercial OCT systems achieve axial resolutions of a few micrometers at normal incidence. XCT, the EUV predecessor of SXCT, typically employs a spectral bandwidth of 70 eV, theoretically enabling resolutions of approximately 10 nm at normal incidence. In practice, XCT is often operated at an incidence angle of 15°, which only slightly reduces the resolution. SXCT, in contrast, uses a broader bandwidth of 230 eV, theoretically allowing axial resolutions of around 3 nm at normal incidence. However, due to low reflectivities in this spectral range, a significantly higher angle of incidence of 72° was chosen in our experiment, resulting in a theoretical resolution limit of approximately 10 nm. The reflectivity curves were calculated for representative materials at the central energies of XCT (65 eV, SiO₂) and SXCT (410 eV, Al₂O₃) based on data of Henke et al.³⁶