Limitations of focused ion beam nanomachining

In this article, some limitations of the processing of structures with dimensions in the nanometer range by focused ion beams will be discussed. In order to enable exact depth control of nanometer structures, the effective sputter yield of silicon was determined as function of the ion dose. At ion doses below ${10}^{16} {\mathrm{cm}}^{\mathrm{-2}},$ the effective sputter yield is not constant and the volume of the area processed increases due to the implantation of ions. Material removal can be measured for doses above ${\text{2×10}}^{16} {\mathrm{cm}}^{\mathrm{-2}}$ and it reaches equilibrium for doses of about ${\text{3×10}}^{17} {\mathrm{cm}}^{\mathrm{-2}}.$ This dose dependence of the effective sputter yield becomes especially effective in beam tail regions with low ion intensity. The shape of nanostructures is further determined by combining the beam shape and the angle dependence of the sputter yield which was experimentally determined. Using this approach with a Gaussian beam shape, a comparison of simulated and measured sidewall angles has shown good agreement for trench structures. Only sidewall regions close to the surface and to the bottom of deep structures show slight deviations. At the surface, non-Gaussian beam tails lead to unintentional sputtering at the corners of the processed area. At the bottom, forward scattered ions lead to higher sputter erosion.