Comparative study of resolution limiting factors in electron beam lithography using the edge roughness evaluation method

Abstract

Resolution capabilities of a Nikon 100 kV experimental column of electron beam (EB) projection lithography, a proof of concept column of low energy electron beam proximity projection lithography (LEEPL), and a variable-shaped EB direct writer were compared using a negative-type chemically amplified resist. Line edge roughness and resolution limit of an isolated line were 2.8 and 40 nm for the Nikon experimental column, and were 4.9 and 52 nm for the EB direct writer at the resist thickness of 250 nm, while they were 10.7 and 61 nm for LEEPL, and were 9.6 and 57 nm for the EB direct writer at the resist thickness of 70 nm. Quantitative influences of resolution limiting factors such as electron-optical beam blur and resist performance were analyzed using the edge roughness evaluation method. The analysis shows that the Nikon experimental column has an ability to resolve 35 nm patterns if resolution performance of the resist is 10% improved. In the high-throughput EB stepper system with a large beam current, it is necessary for resolving 50 nm patterns to keep the electron-optical beam blur under 36 nm. It is known about LEEPL that a necessary and sufficient condition for realizing 50 nm patterns is to develop a 50 nm thick resist having the same resolution performance as the conventional resist dedicated for the use with thicker film thickness, and to obtain a 50 nm pattern mask having the same aperture roughness as the conventional mask with 70 nm patterns.