Influence of sub-100 nm scattering on high-energy electron beam lithography

Electron beam lithography tools have evolved in the direction of higher beam energy in order to achieve high-resolution, fine feature definition. As the beam energy is increased, the "forward" scattering is reduced and the "backscatter" range is increased. Over the years, tools became available data 20 then 50 kV, and now 100 kV operation is common. Operation at higher voltages has several advantages, such as better resolution and process latitude due to reduced forward scattering, and a few disadvantages such as higher dose requirements, substrate heating, and lower contrast for backscatter electron alignment and calibration signals (due to reduced primary electron backscattering generation in thin film). The backscatter range for 100 kV on silicon is about 27 mum compared to 8 mum at 50 kV resulting in different strategies for efficient proximity correction. However, even at 100 kV, scattering in an intermediate range is observed and must be taken into account in order to achieve good linewidth control at the highest resolution. Measurements of the scattering range for both 50 and 100 kV have been made using the point exposure distribution measurement technique [S. A. Rishton and D. P. Kern, J. Vac. Sci. Technol. B 5, 135 (1987)]. For comparison, measurements taken on the same wafer at different voltages show that 50 and 100 kV scattering range functions overlap, after normalizing for the different resist sensitivity, at length scales below 0.5 mum, suggesting a common mechanism, which is independent of the initial electron energy. For thin resists, this suggests that the significant resolution difference between 50 and 100 kV lithography is limited to the "forward" scattering effect as the incident electrons traverse the resist. Extrapolating the scattering function to the approximate beam diameter of 10 nm allows an impulse response function to be numerically determined. The convolution of this function gives reasonably good agreement with dose versus linewidth measurements. (C) 2001 American Vacuum Society.