Advances in High-Throughput Multiple Electron-Beam Lithography
M. Mankos and T. H. P. Chang
Etec Systems, Inc., 26460 Corporate Avenue, Hayward, CA 94545
Summary: High-throughput electron-beam pattern generation provides significant technical
and economic benefits to the semiconductor industry. A multiple electron-beam approach
promises significant improvements in throughput. This paper focuses on two multiple
electron-beam patterning approaches: microcolumn arrays and multisource column.
Multibeam patterning already provides significant technical and economic benefits to mask
patterning in the form of the ALTA® pattern generator. Further benefits can be achieved if feasibility is
proven for multiple electron beams (e-beams) as well. E-beam pattern generation technology1 is capable
of sub-100 nm resolution and precise overlay. The single e-beam pattern generators currently in use, such
as the shaped beam2 and cell projection systems,3 have provided marginal mask throughput, but their
throughput is insufficient for any but the lowest volume direct-write applications. Further, several
fundamental limits of throughput for such single-beam implementations are rapidly approaching.
Throughput is directly proportional to the total current delivered. Due to electron-electron (e-e)
interactions, higher beam currents produce more beam blur and aberrations, which limit lithographic
resolution. In addition, data delivery rate and stage velocity increase with a decrease in minimum feature
size. As resolution requirements are extended to sub-100 nm, economically feasible throughput will be
increasingly difficult to obtain in the shaped beam and cell projection exposure approaches currently used.
Multiple e-beam technology4–9 can offer significant improvements in throughput through the use
of an array of beams, because it minimizes the current required for each beam and eases the data rate
delivery constraints. In some configurations,4 it also lowers the stage velocities. Further, unlike single-
beam approaches, the multibeam approach should produce resolution and precision scalable with
decreasing device dimensions.
2. Microcolumn Approach
In a microcolumn array,4 multiple e-beams are created through the use of an array of closely
spaced, miniature electron-optic columns. Each column contains a single Schottky emission cathode. The
microcolumns typically operate at 1 keV and deliver either a Gaussian or shaped beam. Patterns can be
written with the beam scanned over only a narrow stripe (≤0.1 mm width) with a laser-controlled stage
moving continuously in the orthogonal direction to build up the complete chip pattern. Figure 1a shows
the basic concept of this arrayed microcolumn approach. This system configuration reduces e-e
interactions that occur with beams that are in close proximity to each other. It also minimizes the area
printed by each column and, as a result, reduces the stage velocity.
Considerable progress has been made in advancing the base technologies for the microcolumns.
Electron-optical optimization of the microcolumns has increased the current available to more than 10 nA
while maintaining a spot size sufficiently small for sub-100 nm technology, i.e., approximately 20 to
50 nm. Preliminary 0.1 µm features have been printed with the microcolumn. A simple array of
microcolumns having 20 mm center-to-center spacing has been fabricated and successfully tested.
3. Multisource Approach
In this approach, the electron sources use an array of independently modulated e-beams formed by
laser-driven photocathodes9 to achieve multiple beams. The emitted electrons are collected, collimated,
(Ref. only 3735 EUREM00 MM 000310) 1
and demagnified in a conventional electron-optical column configuration, as shown in Figure 1b, to form
an array of beamlets. This array is scanned across the substrate in a high-speed raster mode by magnetic
and/or electrostatic deflection fields to expose resist-coated wafers. Several photocathode candidates are
currently being developed and evaluated, including metal, e.g., gold, and semiconductor photocathodes,
e.g., the negative electron affinity (NEA) GaAs.5
A 50 kV multisource test bed incorporating a photocathode has been developed and built. The
column allows for a detailed evaluation of both the photocathode sources and the electron optics and is
designed to reduce e-e interactions. The test bed is equipped with a 4-lens magnification stack, which allows
high-resolution imaging of both the photocathode pattern and the demagnified multibeam pattern at the
substrate plane. Photoemission results have been obtained using gold and NEA photocathodes at beam
energies varying from 10 to 50 kV allowing the experimental confirmation of key design parameters.
Figure 1. The basic concept of (a) arrayed microcolumns and (b) a multisource column for high-throughput lithography.
Multiple e-beam approaches offer potential for high-throughput pattern generation for mask
patterning and direct write in the sub-100 nm and below linewidth regimes. Two technologies are
currently being explored: a multimicrocolumn approach and a multiple source approach. The
microcolumns offer better extendibility in throughput because of low e-e interactions and the reduced
requirement for stage velocity. Multiple sources offer better compatibility with existing electron- and
laser-beam technologies. Feasibility demonstrations of both approaches are underway with encouraging
 M. A. McCord, J. Vac. Sci. Technol. B15 (Nov/Dec 1997) 2125.
 H. C. Pfeiffer, et al, J. Vac. Sci. Technol. B11 (Nov/Dec 1993) 2332.
 Y. Nakayama, et al , J. Vac. Sci. Technol. B8 (Nov/Dec 1990) 1836.
 T. H. P. Chang, et al, J. Vac. Sci. Technol. B14 (6) (1996) 3774.
 A. W. Baum, et al, J. Vac. Sci. Technol. B15 (Nov/Dec 1997) 2707.
 T. Newman, W. J. Devore, R. F. W. Pease, J. Vac. Sci. Technol. B1 (Nov/Dec 1983) 999.
 S. Asai, et al, J. Jan. Appl. Phys. 32 (1995) 6012.
 T. Groves, R. A. Kendall, J. Vac. Sci. Technol. B14 (6) (Nov/Dec 1998) 3168.
 M. Mankos, et al, submitted to J. Vac. Sci. Technol., 2000.
The authors wish to acknowledge the technical contributions of their colleagues at Etec Systems,
Inc. Research on multiple e-beam technology at Etec was made with DARPA support under contract
numbers N00019-97-C-2010 and N00019-98-C-0025 of the Department of the Navy.
(Ref. only 3735 EUREM00 MM 000310) 2