Abstract

Several Monte Carlo approaches for modeling electron–solid interaction have been investigated as to their impact on simulation accuracy and performance. An optimum set of models, including the Mott elastic scatter cross-section, the Moller inelastic scatter cross-section, and the continuous slow down approximation with the relativistic Bethe stopping power, have been identified as giving a good agreement with published experimental data while providing better simulation performance than more sophisticated approaches. It has been validated that the Mott cross-section should be used instead of Rutherford's for correct prediction of both electron transmission and backscatter coefficients in the cases of thin films for incident beam energy in the range of 10–100 keV. The use of the Gryzinski cross-section for modeling inelastic scatter did not provide any accuracy improvement because the predicted beam widening in resist was of the same order as given by a more simplistic Moller model. It is also necessary to use the relativistic Bethe stopping power for incident beam energies >30 keV.

Author Keywords: Electron–solid interaction; Backscatter coefficient; Proximity effect; Differential scatter cross-section; Electron stopping power

Corresponding author.