Empirical electron backscatter model for thin resist films on a substrate

Abstract

An electron backscatter model applicable to the exposure of resists in electron lithography is developed. The model is empirical, being based on the direct observation of chemical change produced by backscatter electrons at different accelerating voltages on several substrates. The model is independent of scattering, trajectory, and energy dissipation calculations and is essentially a radial exponential decay of backscatter current density out to the backscatter radius determined by electron range. Predictions of the simple model are compared directly with published results of several authors. The model indicates, in agreement with Nosker, that substrate backscatter is not as deleterious to transverse resolution as predicted by other calculations. The actual numerical agreement among the scattering‐model calculations themselves and this chemical‐effect empirical model is not altogether satisfactory. The empirical model is in better agreement with Nosker at the higher backscatter angles and with Hawryluk et al. at small angles. The contribution of 50‐eV secondary electrons to vinyl ferrocene crosslinking is concluded to be negligible and less than the experimental errors in this model. In the case of highly sensitive resists where the secondaries may be effective a substrate back dose coefficient η_SD must be employed which is about twice the backscatter coefficient η.