Electric fields in Scanning Electron Microscopy simulations

Model improvements to simulate charging in SEM

“…In our previous study [20], the semi-empirical scattering models were designed to get an accurate SE yield value for a semi-infinite surface. However, an accurate SE yield does not guarantee a realistic scattering cloud inside the material, as shown in Figure 2a.…”

“…In the previous study [20], an algorithm based on CSDA was used to model the energy transfer of the primary electrons (PE) [34]. Although CSDA is a good approximation to estimate stopping power (SP) at high energies, it overestimates the SP at very low energies.…”

“…In an earlier paper, we reported a Monte-Carlo simulator, that included charging effects, making use of a multigrid-based electric field solver to decrease the electric field computation time [20]. Combining the electric field solver with a semi-empirical electron-matter interaction model, we were able to simulate an SEM image of an area of ~1.5 µm 2 of grating couplers, shown in Figure 1a, in about 1 hour [20]. However, the shadowing phenomenon (bright-todark transition) on the oxide pads, seen in Figure 1a, could not be reproduced in the simulation (Figure 1c).…”

“…All electrons are traced until their energy is below the surface barrier of approximately 10 eV. Secondly, we addressed issues exclusive to insulators, namely charge redistribution to include effects such as ebeam induced conductivity (EBIC).The simplest approach would be to completely ignore charge redistribution in insulators, as was done in our previous study [20], or only consider catastrophic effects like dielectric breakdown. The other extreme would be modeling from first principles including effects like electron-hole recombination and trapping.…”

“…a) Grating coupler experimental SEM image[20] compared with b) the simulated image without charging, and c) the simulated image with charging taken into account.…”