Nanoscale electron-beam-stimulated processing

Article abstract of DOI:10.1063/1.1565696

Electron beam stimulated deposition and etching were investigated. These processes find application in nanoscale selective processing. During etching of silicon and silicon dioxide, inelastic scattering of the electron beam with the gas was found to take

Full text of DOI:10.1063/1.1565696

Nanoscale electron-beam-stimulated processing
P. D. Rack, S. Randolph, Y. Deng, J. Fowlkes, Y. Choi, and D. C. Joy
Citation: Applied Physics Letters 82, 2326 (2003); doi: 10.1063/1.1565696
View online: http://dx.doi.org/10.1063/1.1565696
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/82/14?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Erratum: “Electron-beam-induced growth of silicon multibranched nanostructures” [Appl. Phys. Lett.87, 113111
(2005)]
Appl. Phys. Lett. 94, 049901 (2009); 10.1063/1.3081382
Effects of developer temperature on electron-beam-exposed hydrogen silsesquioxane resist for ultradense
silicon nanowire fabrication
J. Vac. Sci. Technol. B 25, 2085 (2007); 10.1116/1.2794315
In situ control of the focused-electron-beam-induced deposition process
Appl. Phys. Lett. 83, 4005 (2003); 10.1063/1.1626261
Silicon nitride islands as oxidation masks for the formation of silicon nanopillars
J. Vac. Sci. Technol. A 17, 1294 (1999); 10.1116/1.582100
Nanometer fabrication using selective thermal desorption of SiO 2 induced by focused electron beams and
electron beam interference fringes
J. Vac. Sci. Technol. B 16, 2817 (1998); 10.1116/1.590239
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
155.33.16.124 On: Sat, 29 Nov 2014 13:19:45

APPLIED PHYSICS LETTERS
VOLUME 82, NUMBER 14
7 APRIL 2003
Nanoscale electron-beam-stimulated processing
P. D. Rack,a) S. Randolph, Y. Deng, J. Fowlkes, Y. Choi, and D. C. Joy
Department of Materials Science and Engineering, The University of Tennessee, Knoxville,
Tennessee 37996-2200
͑
Received 27 December 2002; accepted 10 February 2003͒
Electron-beam-stimulated deposition and etching has been investigated as a clean, alternative
method for nanoscale selective processing. Depositions using W(CO) and hydrocarbon sources
6
have yielded efficient and selective electron-beam deposits. Primarily fluorine-based precursors
have been used to etch a variety of materials. Initial results regarding the selective etching of silicon
and silicon dioxide suggest that inelastic scattering of the primary electron beam with the gas occurs
and is more severe at lower beam energies. The etch rate increases linearly with decreasing
electron-beam energy, however, it is not clear if this is due to enhanced primary- or
secondary-electron-stimulated processes. Feature sizes as small as 55 nm have been selectively
processed. © 2003 American Institute of Physics. ͓DOI: 10.1063/1.1565696͔
The ability to manipulate materials at the nanoscale is
critical for the nanotechnology revolution that is occurring.
The intelligent design, fabrication, and repair of nanoscale
devices require techniques to selectively and nanoscopically
deposit and remove material in a controllable fashion. Cur-
rent technologies to selectively deposit or etch microscopic
features utilize ion-beam deposition and etching, laser abla-
tive etching using far-field and near-field optics, and me-
chanical abrasion using a fine microtip. Of these techniques,
focused-ion-beam techniques are probably the most mature
technology that has been extended to application at the
nanoscale. When using an ion beam to stimulate a deposition
or etch process, the gallium ions get implanted into the sub-
strate, which can significantly change the optical, electrical,
or mechanical properties of the substrate. Charging inherent
to the ion–solid interaction also causes proximity effects and
can also lead to so-called ‘‘riverbed effects,’’ which erode
nearby features when the heavy ion beam is scattered and
induces sputtering.
by operating at the ‘‘crossover’’ energy, where the sum of the
secondary and backscattered electron emission coefficients
equals unity, can significantly reduce charging. In addition, if
a low-pressure gas flux is directed onto a surface irradiated
by an electron beam, then the production of positive ions
results, effectively neutralizing the buildup of negative
charge.
To explore the electron-beam-stimulated process, a gas
delivery system was designed and attached to a Hitachi
S-3500N variable pressure ͑SEM͒ ͑VPSEM͒. The VPSEM
has a tungsten hairpin source and is equipped with a back-
scatter detector, an energy dispersive x-ray spectrometer
͑EDS͒, and has a pump system designed to operate in high
vacuum or variable pressure mode ͑up to 0.1–300 Pa͒. The
gas delivery system was designed to deliver up to four gases
to a hypodermic needle for localized gas injection. The in-
jection system is mounted on a wobble stick for three-
dimensional positioning capability.
To date, most of the deposition efforts have focused on
selective tungsten ͑W͒ deposition using a tungsten hexacar-
bonyl ͓W(CO) ͔ source. W(CO) is a solid state source at
Electron-beam-stimulated deposition and etching is con-
ceptually similar to the existing focused-ion-beam approach
and has been shown to be a viable technique for depositing
nanoscopic materials. A variety of materials have been de-
6
6
room temperature and pressure. To deliver the W(CO) gas,
6
the solid source is heated to 45–80 °C, which raises the
vapor pressure enough to deliver ϳ0.1 Pa of gas into the
SEM. EDS measurements were taken in situ to temporally
monitor the deposition process and after the deposition was
completed to quantify the final film deposit. Germanium, as
opposed to silicon, was used as the substrate material be-
cause Si and W have overlapping EDS peak signatures mak-
ing it difficult to determine the composition of the deposit.
Figure 1 shows some initial results of deposition using a
1
posited using a focused electron beam including carbon,
2
3–5
6–7
silicon,8 silicon oxide,9
chromium, gold,
iron,
1
0
11,12
13–21
palladium, platinum,
and tungsten.
Fewer investi-
gators have explored electron-beam etching, however,
2
2,23
24,25
24
photoresist,
silicon,
silicon dioxide,
silicon
2
4
25
nitride, and tantalum/tantalum nitride have been reported.
The main advantage of using an electron beam rather
than an ion beam is reduced contamination. The deleterious
optical, electrical, and mechanical properties that implanted
gallium ions can induce are obviated when using an electron
beam, which is commonly used nondestructively when im-
aging in a scanning electron microscope ͑SEM͒. Another po-
tential advantage of electron-stimulated processes over ion-
beam techniques is the reduced charging effect that can result
in better spatial control of the electron beam. For example,
W(CO) source. A series of 2.5 ␮mϫ2.5 ␮m tungsten boxes
6
were deposited over a range of times from 5–30 min in 5
min increments using a 5 keV beam energy. Figure 1͑b͒ is a
SEM micrograph of the six deposits showing the gradual
thickening of deposits with increasing W(CO) exposure and
6
͑a͒ is the EDS spectra acquired using a probe beam energy of
5
keV, after each run, which shows the germanium substrate
signal decreasing and the W, C, and O peaks increasing with
increasing time. Figure 2 shows an array of ϳ55 and 85 nm
a͒Electronic mail: prack@utk.edu
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
003-6951/2003/82(14)/2326/3/$20.00 2326 © 2003 American Institute of Physics
55.33.16.124 On: Sat, 29 Nov 2014 13:19:45
0
1

Appl. Phys. Lett., Vol. 82, No. 14, 7 April 2003
Rack et al.
2327
FIG. 3. Scanning electron micrograph of an electron-beam-stimulated etch
of an SiO /Si stack.
2
trons emerge from the sample a maximum distance of 5␭
from the beam center where ␭ is the electron inelastic mean-
free path in the specimen. Typical values of ␭ are ϳ1 nm so
the total beam interaction diameter should theoretically be
ϳ60 nm. The etched feature shown in Fig. 3 has a radial
damage pattern 185 nm larger than the prescribed box size
which is ϳ3 times greater than the beam interaction diam-
eter. The discrepancy is believed to be due to primary beam
scattering by the XeF gas.
2
FIG. 1. ͑a͒ EDS spectra and ͑b͒ electron micrograph of electron-beam-
stimulated deposition of W(CO)6 at 5, 10, 15, 20, 25, and 30 min.
Monte Carlo simulation of elastic electron-gas scattering
carbon nanopillars grown from hydrocarbon sources and a
150 nm selectively etched tungsten film.
A significant effort has been directed toward the selec-
tive electron-beam etching of technologically important ma-
terials. Thus far, we have focused on fluorine-based etch
chemistries such as SF and XeF and have etched Si, SiO ,
6
2
2
Si N , Ta, Al, Cr, C, TaN, Cu, low dielectric materials, and
3
4
photoresist. Figure 3 shows an electron-beam-stimulated etch
of a SiO /Si stack that was etched with XeF gas at a total
2
2
background pressure of 0.1 Pa. The localized gas pressure at
the substrate is likely an order of magnitude greater than the
average chamber pressure because of the close proximity of
the expansion nozzle to the substrate (ϳ1 cm).
The lateral dimension of the etched feature shown in Fig.
3
is ϳ620 nm and the vertical depth is 950 nm. The electron
beam was scanned over a square region with a 0.25 ␮m box
edge during the etching process. The diameter of the scan-
ning electron beam in vacuum was ϳ50 nm. Secondary elec-
FIG. 4. ͑a͒ Plot of the silicon etch volume per Coulomb dose vs primary
beam energy and ͑b͒ a plot of the etched silicon feature diameter as a
FIG. 2. Scanning electron micrograph of an array of 55 and 85 nm carbon
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
nanopillars nearby a 150 nm etched tungsten film.
function of primary beam energy.
1
55.33.16.124 On: Sat, 29 Nov 2014 13:19:45

2328
Appl. Phys. Lett., Vol. 82, No. 14, 7 April 2003
Rack et al.
events has been performed, however, elastic scattering angles
are too large to account for the observed broadening. Inelas-
occurring, as most inelastic cross sections are strong func-
tions of energy and decrease with increasing electron energy.
Ϫ3
Ϫ5
tic scattering events with scattering angles of 10 –10 rad
most likely caused the beam broadening that lead to the etch
profile observed. The elastic scattering Monte Carlo simula-
tion is being modified to include small angle inelastic scat-
tering events to better quantify the spatial distribution of
electron flux at the substrate surface.
1
Y. Darici, P. H. Holloway, J. Sebastian, T. Trottier, S. Jones, and J. Ro-
driquez, J. Vac. Sci. Technol. A 17, 692 ͑1999͒.
R. R. Kunz and T. M. Mayer, J. Vac. Sci. Technol. B 6, 1557 ͑1988͒.
K. L. Lee and M. Hatzakis, J. Vac. Sci. Technol. 7, 1941 ͑1989͒.
I. Utke, P. Hoffmann, B. Dwir, K. Leifer, E. Kapon, and P. Doppelt, J. Vac.
2
3
4
Sci. Technol. B 18, 3168 ͑2000͒.
A. Folch, J. Tejada, C. H. Peters, and M. S. Wrighton, Appl. Phys. Lett.
5
A series of experiments were performed to elucidate the
effect of incident beam energy on etch rate and damage ge-
ometry. In this instance, the electron beam was scanning dur-
ing etching over a 0.25 ␮mϫ0.25 ␮m square region and the
beam current used was ϳ1 nA. Figure 4͑a͒ shows a plot of
the silicon etch volume/Coulomb dose as a function of beam
66, 2080 ͑1995͒.
6
R. R. Kunz, T. E. Allen, and T. M. Mayer, J. Vac. Sci. Technol. B 5, 1427
͑
1987͒.
7
8
9
R. R. Kunz and T. M. Mayer, Appl. Phys. Lett. 50, 962 ͑1987͒.
S. Matsui and M. Mito, Appl. Phys. Lett. 53, 1492 ͑1988͒.
S. Lipp, L. Frey, C. Lehrer, B. Frank, E. Demm, S. Pauthner, and H.
Ryssel, J. Vac. Sci. Technol. B 14, 3920 ͑1996͒.
T. J. Stark, T. M. Mayer, D. P. Griffis, and P. E. Russel, J. Vac. Sci.
Technol. B 10, 2685 ͑1992͒.
10
energy with the associated linear regression fit of the data
2
(
R ϭ95%), which shows a linear trend of increased etch
11
O. Yavas, C. Ochiai, M. Takai, A. Hosono, and S. Okuda, Appl. Phys. Lett.
76, 3319 ͑2000͒.
rates at lower beam energies. This trend is indicative of the
expected increase in the inelastic dissociation and ionization
at lower beam energies. However, it is not conclusive
whether the primary electrons or the secondary electrons are
the dominant contributor to XeF2 dissociation. Several
groups have observed higher electron-stimulated deposition
rates at lower energy and have suggested that the rate in-
crease is due to a higher secondary electron yield at lower
energy. An integration of the secondary electron energy dis-
tribution from silicon at 3 keV ͑with a total yield of ␦
12
M. Takai, T. Kishimoto, H. Morimoto, Y. K. Park, S. Lipp, C. Lehrer, L.
Frey, H. Ryssel, A. Hosono, and S. Kawabuchi, Microelectron. Eng. 41Õ
42, 453 ͑1998͒.
13
P. C. Hoyle, J. R. A. Cleaver, and H. Ahmed, Appl. Phys. Lett. 64, 1448
͑1994͒.
H. W. P. Koops, R. Weiel, D. P. Kern, and T. H. Baum, J. Vac. Sci.
Technol. B 6, 477 ͑1988͒.
14
15
P. C. Hoyle, J. R. A. Cleaver, and H. Ahmed, J. Vac. Sci. Technol. B 14,
662 ͑1996͒.
1
1
6
7
K. T. Kohlmann-von Platen, J. Chiebek, M. Weiss, K. Reimer, H. Oertel,
and W. H. Brunger, J. Vac. Sci. Technol. B 11, 2219 ͑1993͒.
R. B. Jackson and J. S. Foord, Appl. Phys. Lett. 49, 196 ͑1986͒.
M. Komuro and H. Hiroshima, Microelectron. Eng. 35, 273 ͑1997͒.
N. A. Kislov, I. I. Khodos, E. D. Ivanov, and J. Barthel, Scanning 18, 114
͑1996͒.
ϭ0.37) and an ionization cross section of SF reveals that
18
6
1
2
9
0
the ratio of ionization events caused by the primary beam
relative to the secondary electrons is ϳ2.5/1. A more-
detailed analysis including dissociation reactions is needed to
fully account for the all of the species ͑ions and radicals͒ that
are participating in the etching reaction. Figure 4͑b͒ is a plot
of the etched feature diameter and the associated linear re-
K. T. Kohlmann, M. Thiemann, and W. H. Brunger, Microelectron. Eng.
13, 279 ͑1991͒.
21
S. Matsui and K. Mori, J. Vac. Sci. Technol. B 4, 299 ͑1986͒.
K. T. Kohlmann-von Platen and W. H. Brunger, J. Vac. Sci. Technol. B 14,
22
4262 ͑1996͒.
23
2
S. Matsui, T. Ichihashi, and M. Mito, J. Vac. Sci. Technol. 7, 1182 ͑1989͒;
R. R. Kunz and T. M. Mayer, J. Vac. Sci. Technol. B 5, 427 ͑1987͒.
J. W. Coburn and H. F. Winters, J. Appl. Phys. 50, 3189 ͑1979͒.
P. E. Russell ͑private communications͒.
gression fit of the data (R ϭ98%) and reveals that the ef-
fective spot size increased with decreasing beam energy. This
trend agrees with speculation that inelastic gas scattering is
24
25
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
55.33.16.124 On: Sat, 29 Nov 2014 13:19:45
1