Using a neural network to proximity correct patterns written with a Cambridge electron beam microfabricator 10.5 lithography system

This letter describes our initial results of using a theoretical determination of the proximity function and an adaptively trained neural network to proximity correct patterns written on a Cambridge electron beam lithography system. The methods described are complete and may be applied to any electron beam exposure system that can modify the dose during exposure. The patterns produced in resist show the effects of proximity correction versus noncorrected patterns.