In 45nm node and beyond with hyper NA lithography, mask topography effect is not ignorable and mask CD bias
impacts printing performance such as MEEF or exposure latitude. In that situation, 3D simulation is required for precise
evaluation of printing performance and the accuracy of 3D mask model on simulation is a key issue. Verification of 3D
mask model by diffraction intensity measurement with AIMSTM45-193i was discussed in our previous works. Through
the verification, though real mask successfully creates effective or simulated diffractions, CD on 3D mask model on
simulation was different to that on AIMSTM result which was measured by CD-SEM. Therefore, purpose of this work is
to analyze the cause of CD differences through AIMSTM diffraction intensity evaluation in various conditions (mask
material, pattern pitch, mask CD bias and mask CD-SEM system). Furthermore, lithography equivalent CD is proposed
as width of "ideal" mask shape.
As a result achieved from the experiments, constant CD shift was successfully observed at hp40-70nm L/S pattern with
varied bias for both 6% EAPSM and Binary masks. It can be said that mask topography difference related to mask
material and pattern dimensions has not been observed. On the other hand, the value of CD shift was smaller on the
condition of newer generation CD-SEM measurement.
Other result achieved from further discussion and analysis, cause of the CD difference was explained using simple SEM
image simulation. The CD difference was mainly changed by electron beam size factor, and it was stable with side wall
angle in the range of 80 to 90 degree if the middle CD, which is the width of 3D model defined at the half height of the
mask film's thickness, is constant. Since side wall angles on actual masks are nearly 90 degree, lithography equivalent
CD could be measured by CD-SEM with constant offset.