Lq. Han et al., Performance investigation of Coulomb interaction-limited high through put electron beam lithography based on empirical modeling, J VAC SCI B, 16(6), 1998, pp. 3215-3220
Multibeam direct writing and projection strategies have been proposed for h
igh throughput electron beam lithography; however, the large beam current r
equired may cause severe beam blur due to the stochastic Coulomb interactio
n effects between the electrons. From viewpoint of both concept proof and p
ractical system design, evaluation of:the fundamental system limits and the
optimum performance in a large range of multiple system parameter variable
space is critical for such-high throughput tools. Based on the response su
rface design approach, well-fitted empirical models of image beam size as a
function of various system parameters have been extracted from Monte Carlo
simulations of electron interactions for different proposed multi-beam sys
tems. The blurred imaging beam size is examined in a wide range of system p
arameters: beam currents from 1 to 100 mu A, acceleration voltages;from 25
to 100 kV, column lengths from 1 to 100 cm, field sizes from 0.5 to 2 mm, a
nd demagnification from 4 to 16. Two general tendencies for such systems we
re observed: the nonuniformity of image defocusing due to the inhomogeneous
interbeamlet space charge effect at large field sizes, and the high sensit
ivity of maximum beam current to column length at a fixed image beam size.
Finally, the throughput versus acceleration voltage is evaluated under diff
erent assumptions of resolution, resist thickness, and column length. It is
concluded that to achieve the throughput required by manufacturing beyond
0.1 mu m resolution, very aggressive design of a short column along with th
in resist will be required. (C) 1998 American Vacuum Society. Society. [S07
34-211X(98)13906-9].