Determination of proximity effect correction parameters for 0.1 μm electron-beam lithography

Abstract

Proximity correction parameters for electron-beam lithography are generally determined by curve fitting an analytical multiple gaussian expression to Monte Carlo or experimental radial absorbed energy distributions. However, the curve fitting is not a well posed procedure with the result that different methods of curve fitting can result in significantly different proximity correction parameters. We will investigate the critical issues in implementing a multiple gaussian model for accurate proximity correction for 0.1 μm electron-beam lithography. In particular, we will discuss several methods of fitting multiple gaussian expressions to the absorbed energy distribution and examine the technique used in a previous paper1 to fit a three gaussian expression for 0.1 μm lithography on Si and GaAs substrates. It will be shown that it is often necessary to compensate for various process conditions by adjusting the proximity correction parameters. Curve fitting to experimental absorbed energy distributions will be considered and a normalized gaussian expression will be found to give the most consistent fits to the data. The wide variation in reported proximity correction parameters in the literature will be explained. The use of more than three gaussians to exactly model the absorbed energy distribution will also be discussed. © 1990.