Distortion of critical dimension (CD) in electron beam (e-beam) lithography
due to resist heating has become a major error factor in high-throughput m
askmaking. Resist heating was examined using 50 kV raster and vector writin
g strategies for a single pass and for four passes. All writing and process
ing conditions (e-beam column, flash size, resist, development, writing thr
oughput, etc.) were equal except for beam deflection. A 13.8% effective dos
e change due to resist heating was found in the worst raster exposure condi
tions, while in vector exposure, this value was greater than 130%. In four-
pass printing, dose change was less than 1% for raster writing and 16.8% Fo
r vector writing. Simulations of temperature rise were done using TEMPTATIO
N (temperature simulation) software. The temperature rise was found to be 3
8 degrees C for raster exposure but greater than 220 degrees C with vector
exposure for the writing conditions used. SPR 700 and EER 900 novolac resis
ts showed similar behavior with regard to heating.