Content area
This work concludes our systematic investigations of the spatial analytical resolution at different primary electron energies and specimens with different atomic numbers [1-4]. The aim of the entire study was to determine the qualitative and especially the quantitative analytical spatial resolutions that can actually be achieved in practise. For this purpose, we prepared test specimen by evaporating thin metallic layers (Al, Ag and Au) with various thicknesses on top of non-fluorescent substrates. To determine the depth resolution, the layers were analysed from the top. The measurements were carried out using a field emission electron microprobe, but the results are also applicable to SEM-EDS and SEM-WDS.
This work now evaluates the quantitative analytical depth resolution measured with electron energies from 4 to 15 keV on eight gold layer specimens with different layer thicknesses from 30 to 1,030 nm. By varying the primary electron energy E0, we determined the minimum value of E0 for which the depth of the signal source volume is identical to the layer thickness. This value is defined as the depth resolution, which is identical to the ionisation range of the atomic shell under consideration. With respect to the excitation energy of the characteristic X-rays, the depth resolution for the quantitative element analysis for Au (Mα at 4 kV) is found to be below 50 nm. It should be noted that the applied electron energies fulfil the 2-times overvoltage criteria resulting in reasonable X-ray signals with good signal-to-noise ratio.
The experimental results were cross-checked with simulations and calculations. While we found a good agreement with Monte Carlo simulations, the well-known Castaing-formula predicts smaller values for all primary electron energies. The deviation increases for decreasing energies. Since we regard test specimens without any disturbing secondary fluorescence, the measurements presented in this paper are very helpful for further improvements of calculations, MC simulations and other approximations.