Self-limiting atomic-layer deposition of Si on SiO2 by alternate supply of Si2H6 and SiCl4

Article abstract of DOI:10.1063/1.1389508

Atomic-layer deposition of Si on SiO₂ with a self-limiting growth mode was achieved at substrate temperatures between 355 and 385 °C by means of alternate supply of Si₂H₆ and SiCl₄ gas sources. The growth rate was saturated at 2 ML per cycle at these temperatures and for Si₂H₆ exposure time over 120 s. The smooth surface (～0.26 nm in arithmetic average roughness) was obtained under the self-limiting condition irrespective of a film thickness up to 6.5 nm.

Full text of DOI:10.1063/1.1389508

Self-limiting atomic-layer deposition of Si on SiO2 by alternate supply of
Si2H6 and SiCl4
Shin Yokoyama, Kenji Ohba, and Anri Nakajima
Citation: Appl. Phys. Lett. 79, 617 (2001); doi: 10.1063/1.1389508
View online: http://dx.doi.org/10.1063/1.1389508
View Table of Contents: http://apl.aip.org/resource/1/APPLAB/v79/i5
Published by the AIP Publishing LLC.
Additional information on Appl. Phys. Lett.
Journal Homepage: http://apl.aip.org/
Journal Information: http://apl.aip.org/about/about_the_journal
Top downloads: http://apl.aip.org/features/most_downloaded
Information for Authors: http://apl.aip.org/authors
Downloaded 01 Sep 2013 to 128.103.149.52. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://apl.aip.org/about/rights_and_permissions

APPLIED PHYSICS LETTERS
VOLUME 79, NUMBER 5
30 JULY 2001
Self-limiting atomic-layer deposition of Si on SiO by alternate supply
2
of Si H and SiCl
2
6
4
Shin Yokoyama,a) Kenji Ohba, and Anri Nakajima
Research Center for Nanodevices and Systems, Hiroshima University, 1-4-2 Kagamiyama,
Higashi-Hiroshima, Hiroshima 739-8527, Japan
͑
Received 10 January 2001; accepted for publication 4 June 2001͒
Atomic-layer deposition of Si on SiO with a self-limiting growth mode was achieved at substrate
2
temperatures between 355 and 385 °C by means of alternate supply of Si H and SiCl gas sources.
2
6
4
The growth rate was saturated at 2 ML per cycle at these temperatures and for Si H exposure time
2
6
over 120 s. The smooth surface ͑ϳ0.26 nm in arithmetic average roughness͒ was obtained under the
self-limiting condition irrespective of a film thickness up to 6.5 nm. © 2001 American Institute of
Physics. ͓DOI: 10.1063/1.1389508͔
Atomic-layer-controlled deposition ͑ALD͒ has recently
been attracting a great deal of attention as the size of the
electronic devices decreases. Especially for gate insulators
Torr, SiCl ͑200 Torr, at a substrate temperature T ϭ375 °C:
4 1
fixed͒ and Si H ͑20%, He base, 200 Torr, T ϭ345–420 °C͒
2
6
2
gases are alternately supplied. For the SiCl exposure, we
4
1
2
with a high dielectric constant such as Al O , ZrO , and
employed the same condition used in the self-limiting
atomic-layer deposition of silicon nitride by alternate supply
2
3
2
3
HfO , the ALD process is recognized as a practical method
2
1
3
because of its excellent thickness uniformity, low growth
temperature, and less particle generation.
of SiCl and NH , because the SiCl adsorption was re-
4 3 4
ported to be saturated at ϳ1 ML at this condition. The typi-
cal growth cycle was 20. The film thickness was measured
by the ellipsometry.
Besides the dielectric films, the ALD of semiconducting
Si films on SiO is very useful, for example, for quantum
2
devices, in which a small Si island with precisely controlled
size is necessary to be sandwiched in the insulating films.4
This technology will be useful for the ultra-large-scale inte-
grated ͑ULSI͒ circuits, for example, to fabricate a precisely
size-controlled floating gate memory and accurate resistors.
It is also applicable to fabricate a silicon on insulator sub-
strate having an extremely flat thin Si layer.
The growth rate is shown in Fig. 2 as a function of
reciprocal substrate temperature at Si2H6 exposure. The Si
growth rate for the exposure of only Si2H6 ͑the same condi-
tion except for no supply of SiCl4͒ is also plotted in Fig. 2.
For the alternate gas supply of SiCl4 and Si2H6, the growth
rate is nearly constant ͑ϳ2 ML/cycle͒ in the temperature
range between 355 and 385 °C. On the other hand, for the
gas exposure of only Si2H6, the growth rate exponentially
increases with the temperature and the activation energy of
ϳ2.0 eV is obtained from Fig. 2, which coincides with that
for the thermal dissociation of Si2H6 into SiH4 and SiH2 .14
Without the exposure of Si2H6, i.e., only the SiCl4 exposure,
the film growth did not take place. Therefore, the HCl de-
sorption reaction of SiCl4 with the surface H–Si bond,
and/or HCl desorption reaction of Si2H6 with the surface
Cl–Si bond probably cause the self-limiting growth of ϳ2
ML/cycle in the temperature range 355–385 °C. Namely,
Although there are many reports on the atomic-layer ep-
5
–8
itaxy ͑ALE͒ of Si, there are only a few reports concerning
the ALD of Si on SiO2.9 Tanaka et al. reported the ALD
,10
9,10
of Si on SiO by cryogenic laser-induced deposition using
2
Si H . They obtained a self-limiting growth mode, in which
2
6
the growth rate is automatically adjusted to a monolayer per
cycle, at substrate temperatures Ϫ26 to Ϫ49 °C. However, it
may be difficult to obtain a high-purity Si film because the
residual impurity gases in the vacuum chamber will be easily
adsorbed on the cooled substrate. In this letter, we have
achieved atomic-layer deposition of Si on SiO with a self-
2
limiting growth mode at more practical substrate tempera-
tures by means of alternate supply of SiCl and Si H . The
4
2
6
growth rate is nearly constant ͑ϳ2 ML/cycle͒ at temperatures
between 355 and 385 °C, and a very smooth surface is ob-
tained irrespective of the film thickness.
The time sequence for temperature and gas supply is
shown in Fig. 1. The ALD system consists of a quartz cham-
3
ber ͑volume of ϳ100 cm ͒, a turbomolecular pump ͑1500
1
1
l/s͒, and computer-controlled halogen lamp and gas valves.
The substrate is OH-terminated SiO2 ,12 which was obtained
by dilute HF treatment of thermally oxidized p-Si͑100͒ at
FIG. 1. Typical time sequence of gas supply and substrate temperature
1
000 °C, followed by a short deionized water rinse ͑1 min͒
control. T1 and T2 indicate the substrate temperature during SiCl exposure
4
Ϫ7
and spin dry. After evacuating the chamber into Ͻ2ϫ10
and Si2H6 exposure, respectively. The substrate temperature was computer-
controlled synchronized with the gas supply. The vacuum evacuation time
between SiCl and Si H exposures is 90 s. The pressure in the chamber just
4
2
6
a͒Electronic mail: yokoyama@sxsys.hiroshima-u.ac.jp
before each gas supply is ϳ2ϫ10 Torr.
Ϫ4
0003-6951/2001/79(5)/617/3/$18.00
617
© 2001 American Institute of Physics
Downloaded 01 Sep 2013 to 128.103.149.52. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://apl.aip.org/about/rights_and_permissions

618
Appl. Phys. Lett., Vol. 79, No. 5, 30 July 2001
Yokoyama, Ohba, and Nakajima
FIG. 4. Surface roughness vs substrate temperature at Si H exposure. The
2
6
growth cycle is 20. The SiCl4 and Si H6 exposure times are 4 and 2 min,
2
FIG. 2. Growth rate of Si as a function of substrate temperature at Si H6
respectively.
2
exposure. The SiCl4 and Si H exposures times are 4 and 2 min, respec-
2
6
tively. Two monolayer thicknesses of 0.27 nm for Si͑100͒ and 0.31 nm for
Si͑111͒ are indicated in the figure.
5
rate. While in the Si Cl /Si H case, a low Si H gas pres-
2
6
2
6
2
6
Ϫ3
sure of 2ϫ10 Torr was sufficiently enough to obtain a
practically large growth rate because of the high substrate
each HCl desorption reaction automatically stops when 1 ML
of Si covers the surface. Here, the Cl- and/or H-terminated
surface acts as a surfactant to operate the self-limiting
temperature. On the other hand, concerning the SiCl expo-
4
sure condition, the monolayer adsorption of SiCl on the
4
OH-terminated SiO2 surface18 and on the NH-terminated
1
5
mechanism. While at higher temperatures ͑Ͼ400 °C͒ the
silicon nitride surface13 has been indirectly proven, respec-
dissociation of Si H becomes predominant and the growth
2
6
1
8
tively, from the ALD of SiO using a SiCl /H O system
rate increases independent on the SiCl exposure.
2
4
2
4
and from the ALD of silicon nitride using a SiCl /NH
The deposition rate for the SiCl /Si H alternate supply
4
3
4
2
6
system.13 In these cases, a relatively high gas pressure of
at a substrate temperature of 355 °C ͑at Si H exposure͒ is
2
6
1
–10 Torr ͑Ref. 18͒ or 200 Torr ͑Ref. 13͒ and a relatively
shown in Fig. 3 as a function of Si H exposure time. The
2
6
low temperature ͓427 °C ͑Ref. 18͒ or 375 °C ͑Ref. 13͔͒ con-
deposition rate saturates at a roughly 2 ML/cycle for the
Si H exposure time longer than 120 s.
ditions were used. Therefore, we have employed the same
2
6
1
3
SiCl exposure condition as the SiCl /NH system. It is
Here, we discuss the reason for the success in the self-
4
4
3
speculated from the rate saturation of the ϳ2 ML/cycle in
limiting growth in our method. In order to obtain the self-
limiting growth mode, an alternate supply of chloride and
hydride gas sources has been usually employed for the ALD
Figs. 2 and 3 that the SiCl adsorption is also saturated at ϳ1
4
ML on the SiH-terminated surface at the condition used.
Figure 4 shows the surface roughness measured by
atomic-force microscopy as a function of substrate tempera-
ture at Si H exposure. Below 395 °C the surface is smooth
5
,8,16–18
5
or ALE.
Gates et al. reported an ALE of Si using
Si Cl and Si H , however, the self-limiting growth mode
2
6
2
6
was not achieved. The essential reason for our success in the
self-limiting growth for the SiCl /Si H system is probably
2
6
͑
ϳ0.26 nm in arithmetic average roughness: R ͒, while it
a
4
2
6
becomes rough at temperatures over 395 °C, possibly due to
an employment of a relatively low substrate temperature
355–385 °C͒ at Si H exposure. While in the Si Cl /Si H
5
Si nucleation. Figure 4 indicates the growth mode change
͑
2
6
2
6
2
6
1
5
5
from the layer by layer ͑Frank–Van der Merwe ͒ mode to
island growth ͑Volmer–Weber19 or Stranski–Krastanov ͒
mode at high temperatures. The surface roughness is also
case, the substrate temperature was relatively high ͑465 °C͒,
which may cause continuous Si deposition through thermal
dissociation of Si H at the substrate surface. In compensa-
20
2
6
tion for using the low substrate temperature, we employed a
high gas pressure of 200 Torr in order to gain the growth
FIG. 5. Surface roughness vs film thickness. The substrate temperature at
Si H exposure is fixed at 355 °C. The gas exposure times are the same as in
2
6
FIG. 3. Deposition rate of Si vs Si H exposure time at a substrate tempera-
Fig. 4. It was confirmed that the film thickness was proportional to the
growth cycle.
2
6
ture of 355 °C. The SiCl exposure time is fixed at 4 min.
4
Downloaded 01 Sep 2013 to 128.103.149.52. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://apl.aip.org/about/rights_and_permissions

Appl. Phys. Lett., Vol. 79, No. 5, 30 July 2001
Yokoyama, Ohba, and Nakajima
619
4
5
plotted as a function of the film thickness in Fig. 5. It is
recognized that the surface is maintained smooth up to a film
thickness of 6.5 nm as far as the self-limiting condition is
employed.
For example, see A. Nakajima, T. Futatsugi, K. Kosemura, T. Fukano, and
N. Yokoyama, Appl. Phys. Lett. 71, 353 ͑1997͒.
S. M. Gates, D. D. Koleske, J. R. Heath, and M. Copel, Appl. Phys. Lett.
62, 510 ͑1993͒.
6
7
8
9
J. Murota, M. Sakuraba, and S. Ono, Appl. Phys. Lett. 62, 2353 ͑1993͒.
D. D. Koleske and S. M. Gates, Appl. Phys. Lett. 64, 884 ͑1994͒.
D. D. Koleske and S. M. Gates, J. Appl. Phys. 76, 1615 ͑1995͒.
T. Tanaka, T. Fukuda, Y. Nagasawa, S. Miyazaki, and M. Hirose, Proceed-
ings of 1989 Dry Process Symposium Tokyo, 1989 ͑The Institute of Elec-
trical Engineers of Japan, Tokyo, 1989͒, p. 103.
T. Tanaka, T. Fukuda, Y. Nagasawa, S. Miyazaki, and M. Hirose, Appl.
Phys. Lett. 56, 1445 ͑1990͒.
K. Ooba, Y. Nakashima, A. Nakajima, and S. Yokoyama, Ext. Abst. Conf.
Solid State Devices and Mater., Hiroshima, 1998 ͑Business Center for
academic Societies Japan, Tokyo, 1998͒, p. 22.
S. Miyazaki, Y. Hamamoto, E. Yoshida, M. Ikeda, and M. Hirose, Thin
Solid Films 369, 55 ͑2000͒.
S. Yokoyama, Y. Nakashima, and K. Ooba, J. Korean Phys. Soc. 35, S71
͑1999͒.
M. Browery and J. H. Purnell, Proc. R. Soc. London 321, 341 ͑1971͒.
M. Copel, M. C. Reuter, M. Horn von Hoegen, and R. M. Tromp, Phys.
Rev. B 42, 11682 ͑1990͒.
In conclusion, we have demonstrated the self-limiting
growth of Si on SiO at practical growth temperatures 355–
2
385 °C using chloride/hydride gas sources. A smooth surface
was obtained under the self-limiting growth condition. The
epitaxial Si growth with the self-limiting mechanism would
be expected under this growth condition. This technique will
help in the realization of sophisticated ULSI and quantum
devices.
10
11
12
This work has been supported by a Grant-in-Aid for Sci-
entific Research ͑B͒ from the Ministry of Education, Sci-
ence, Sports and Culture, the Japanese Government ͑Project
No. 11450125͒, and the Core Research for Evolutional Sci-
ence and Technology ͑CREST͒ of the Japan Science and
Technology Corporation ͑JST͒.
13
1
1
4
5
16
J. Nishizawa, K. Aoki, S. Suzuki, and K. Kikuchi, J. Electrochem. Soc.
1
37, 1898 ͑1990͒.
1
17
S. M. George, O. Sneh, A. C. Dillon, M. L. Wise, A. W. Ott, L. A. Okada,
and J. D. Way, Appl. Surf. Sci. 82Õ83, 460 ͑1994͒.
M. Copel, M. Gribelyuk, and E. Gusev, Appl. Phys. Lett. 76, 436 ͑2000͒.
M. Pessa, R. Makela, and T. Suntola, Appl. Phys. Lett. 38, 131 ͑1981͒.
J. W. Klaus, A. W. Ott, J. M. Johnson, and S. M. George, Appl. Phys. Lett.
70, 1092 ͑1997͒.
H. Kawabata, H. Ueba, and C. Tatsuyama, J. Appl. Phys. 66, 634 ͑1989͒.
E. Bauer and H. Poppa, Thin Solid Films 12, 167 ͑1972͒.
18
2
3
19
K. Kukli, J. Ihanus, M. Ritala, and M. Leskela, Appl. Phys. Lett. 68, 3737
20
͑1996͒.
Downloaded 01 Sep 2013 to 128.103.149.52. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://apl.aip.org/about/rights_and_permissions