sensitivity, at length scales below 0.5pm, suggesting a common mechanism,
which is independent of the initial electron energy. For thin resists, this suggests
that the significant resolution difference between 50KV and 100KV 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
10nm allows an impulse response function to be numerically determined. The
convolution of this function gives reasonably good agreement with dose versus
line-width measurements.
Introduction
In electron beam lithography, the incident beam is scattered by the resist and the
substrate resulting in unwanted dose being deposited in areas outside of the
desired exposure region. Normally, this scattering is discussed in terms of
"backscatter" which is the result of a large number of electron scattering events
and where resist can be exposed quite some distance from the electron beam
position and "forward" scattering which is the result of a small number of
scattering events as the electron beam passes through the resist. In thicker resist,
forward scattering will have a larger range than thin resist. For high accelerating
voltage systems, forward scattering has a significantly smaller range than
backscatter. In reality, scattering is not so neatly divided into two distinct regimes.
A practical technique for investigating scattering is the point exposure distribution
measurement technique as described by Rishton and Kern.1 In this technique, a
series of dot point exposures with exponentially increasing dose over many orders
of magnitude is carried out using a high-contrast resist as the scattering detector.