Effect of hydrogen plasma precleaning on the removal of interfacial amorphous layer in the chemical vapor deposition of microcrystalline silicon films on silicon oxide surface

Article abstract of DOI:10.1063/1.115864

Microcrystalline silicon films were deposited on SiO₂ substrates and the effect on in situ hydrogen plasma precleaning was investigated. In situ hydrogen plasma was found effective in removing oxygen and carbon impurities and also found to indu

Full text of DOI:10.1063/1.115864

Effect of hydrogen plasma precleaning on the removal of interfacial amorphous layer in
the chemical vapor deposition of microcrystalline silicon films on silicon oxide surface
YoungBae Park and ShiWoo Rhee
Citation: Applied Physics Letters 68, 2219 (1996); doi: 10.1063/1.115864
View online: http://dx.doi.org/10.1063/1.115864
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/68/16?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Very lowroughness diamond film deposition using a surfacewave sustained plasma
J. Appl. Phys. 80, 6013 (1996); 10.1063/1.363600
Substrate dependence of initial growth of microcrystalline silicon in plasmaenhanced chemical vapor deposition
J. Appl. Phys. 80, 6061 (1996); 10.1063/1.363565
Rapid solid phase crystallization of nanocrystalline silicon deposited by electron cyclotron plasma chemical vapor
deposition
Appl. Phys. Lett. 69, 1873 (1996); 10.1063/1.117462
Nucleation and bulk film growth kinetics of nanocrystalline diamond prepared by microwave plasmaenhanced
chemical vapor deposition on silicon substrates
Appl. Phys. Lett. 69, 1716 (1996); 10.1063/1.118007
Effects of hydrogen surface pretreatment of silicon dioxide on the nucleation and surface roughness of
polycrystalline silicon films prepared by rapid thermal chemical vapor deposition
Appl. Phys. Lett. 69, 485 (1996); 10.1063/1.118148
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:
130.239.20.174 On: Sat, 22 Nov 2014 11:21:01

Effect of hydrogen plasma precleaning on the removal of interfacial
amorphous layer in the chemical vapor deposition of microcrystalline silicon
films on silicon oxide surface
Young-Bae Park and Shi-Woo Rheea)
Laboratory for Advanced Materials Processing (LAMP), Department of Chemical Engineering, Pohang
University of Science and Technology (POSTECH), 790-784 Pohang, Korea
͑
Received 18 December 1995; accepted for publication 9 February 1996͒
Microcrystalline silicon͑␮c-Si͒ film deposited on silicon oxide in a remote plasma enhanced
chemical vapor deposition ͑RPECVD͒ with disilane (Si H ) and silicon tetrafluoride (SiF ) has
2
6
4
been investigated. It was found that in situ hydrogen plasma cleaning of the substrate prior to
deposition is effective to reduce the interfacial amorphous transition region. It is believed that
hydrogen plasma cleaning generated adsorption and nucleation sites by breaking weak Si–O and
Si–Si bonds and also removed oxygen/carbon impurity. Surface roughening was observed from the
hydrogen plasma precleaning which helped nucleation and crystallization at the initial stage of the
growth. © 1996 American Institute of Physics. ͓S0003-6951͑96͒01016-3͔
Recently, microcrystalline silicon films ͑␮c-Si͒ have
been used for thin-film transistors ͑TFTs͒ in the field of ac-
tive matrix liquid crystal displays ͑AMLCDs͒. The use of
crystalline silicon as an active layer instead of amorphous
silicon has drawn much attention because its electron mobil-
ity is much higher than that of a-Si and better performance
can be achieved with poly-Si TFT.1 In AMLCD applica-
tion, low-temperature processing below 600 °C is desirable
to reduce the manufacturing cost by utilizing low cost glass
substrate. However, it is relatively difficult to grow crystal-
line films because adatom mobility is low and incorporation
of impurities, especially oxygen is more likely.
chemical cleaning effects of fluorine and crystalline phase
can be grown at low temperatures.1
3,14
In a remote plasma enhanced CVD reactor, hydrogen
(H ) is introduced into an alumina tube and excited by in-
2
ductively coupled rf ͑13.56 MHz͒ plasma and flown into the
deposition zone. This plasma is used for the SiO substrate
2
–3
precleaning. For deposition, disilane (Si H ) and silicon tet-
2
6
rafluoride (SiF ) are introduced into the region 10 cm down-
4
stream of the plasma. These reactant gases are mixed with
excited hydrogen and flown toward the substrate 20 cm
downstream of the plasma. Thermally oxidized 4 in. silicon
wafers with an oxide thickness of 100 mm were used as
substrates. For in situ hydrogen plasma cleaning, the pres-
sure was 0.3 Torr, temperature was 300 °C, plasma power
and hydrogen flow rate were 100 W and 100 sccm, respec-
tively. On the other hand, for ␮c-Si deposition, reactor pres-
sure was 0.4 Torr, deposition temperature was 430 °C,
In situ plasma cleaning prior to silicon epitaxy has been
studied to remove remaining surface contaminants such as
oxygen and carbon from single-crystal silicon substrate sur-
face at low temperatures ͑Ͻ600 °C͒. Various sources such as
4
–6
radio frequency remote plasma
resonance ͑ECR͒
and electron cyclotron
7
–11
plasma excited argon, helium, and hy-
plasma power and SiF flow rates were 60 W and 0–30
4
drogen have been reported. Hydrogen plasma cleaning is
from a chemical etching by hydrogen atoms and ions and it
has been used as a substitute for inert ͑Ar and He͒ plasma
cleaning to avoid physical damage.9
sccm, respectively. The deposition rate was in the range of
3
–10 nm/min.
The surface roughness of SiO substrate and deposited
2
,11
Si film was observed using atomic force microscopy ͑AFM͒.
The root-mean-square ͑rms͒ value of surface roughness was
obtained by averaging ten different areas. The interfacial
composition was examined by Auger electron spectroscopy
Semiconductor active layers in thin-film transistor appli-
cations are usually grown on a substrate which is not crys-
talline. The effects of surface roughness and contamination
of amorphous substrate on the growth and microstructure of
the active layer are not fully understood. It has been thought
that the surface roughness may affect the diffusion of the
deposition precursor and relaxation of the Si network during
͑AES͒ at a voltage of 3.5 keV. To observe the microstructure
of the silicon film on SiO , transmission electron microscopy
2
͑TEM͒ was used. Cross-sectional specimens for TEM were
first glued face to face, cut, then ground, dimpled, and
thinned by ion milling. TEM observations were conducted
using a transmission electron microscopy operating at 100
kV.
We have performed in situ hydrogen plasma cleaning
and investigated the surface roughness of the substrate sur-
face. AFM images in Fig. 1 show the surface roughening as
a function of in situ hydrogen plasma cleaning time. The rms
value of surface roughness of thermal oxide is 0.2–0.49 nm
without any treatment. The rms of oxide surface is increased
to 1.65 nm for 10 min and 2.2 nm for 30 min hydrogen
plasma cleaning, respectively. The creation of free-surface
1
2
film growth.
In this research, microcrystalline silicon film deposited
on silicon oxide in a remote plasma enhanced chemical va-
por deposition ͑RPECVD͒ with disilane (Si H ) and silicon
2
6
tetrafluoride (SiF ) has been investigated. We have investi-
4
gated the effect of in situ hydrogen plasma cleaning of the
SiO substrate on the initial nucleation and crystal growth.
2
We added SiF because it is generally known that oxygen
4
and other impurity incorporation can be reduced from in situ
a͒Electronic mail: srhee@vision.postech.ac.kr
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:
Appl. Phys. Lett. 68 (16), 15 April 1996 0003-6951/96/68(16)/2219/3/$10.00 © 1996 American Institute of Physics 2219
30.239.20.174 On: Sat, 22 Nov 2014 11:21:01
1

FIG. 2. Cross-sectional TEM images and their selective area diffraction
SAD͒ pattern of Si films deposited with Si H /SiF /H ϭ0.1/30/100 at a
͑
2
6
4
2
temperature of 430 °C, pressure of 0.4 Torr and plasma power of 60 W. ͑a͒
Bright field image without hydrogen plasma cleaning, ͑b͒ SAD for ͑a͒, ͑c͒
bright field and dark field image with hydrogen plasma cleaning for 5 min
and ͑d͒ SAD for ͑c͒.
FIG. 1. Three-dimensional AFM images of the SiO substrate with hydro-
2
gen plasma cleaning for ͑a͒ 0, ͑b͒ 10, and ͑c͒ 30 min. Plasma power at 100
W, substrate temperature at 300 °C, and pressure at 0.3 Torr.
area and roughness on the initial growing surface was due to
the inhomogeneous etching by hydrogen plasma. Hydrogen
radicals preferentially remove weak Si–Si bond and break
Si–O bond at the top of the SiO surface to create voids and
2
dangling bonds.
Silicon atoms arriving at roughened surface easily ad-
here to the surface sites and are immediately trapped by
voids and dangling bonds on the SiO surface. In fact, in situ
2
hydrogen plasma cleaning prior to deposition increased the
deposition rate by more than 15%. It is also expected that
plasma cleaning will help the film growth with crystalline
character from the early stage of film growth.
Figure 2 shows cross-sectional TEM images. Without in
situ hydrogen plasma cleaning ͓Fig. 2͑a͔͒, the interface re-
FIG. 3. AES depth profile ͑a͒ without and ͑b͒ with hydrogen plasma clean-
ing. Plasma power at 100 W, temperature at 300 °C, pressure at 0.3 Torr and
plasma cleaning for 5 min.
gion of deposited-Si film on amorphous SiO does not show
2
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:
220 Appl. Phys. Lett., Vol. 68, No. 16, 15 April 1996 Y.-B. Park and S.-W. Rhee
30.239.20.174 On: Sat, 22 Nov 2014 11:21:01
2
1

at. % level. It means that the atomic hydrogen reacted with
surface impurities and reaction products were desorbed as a
form of volatile hydrides. It shows that hydrogen plasma
cleaning is effective in removal of impurities.
Figure 4 shows AFM images of silicon films deposited
with or without SiF addition and with in situ hydrogen
4
plasma precleaning. Without SiF addition, the Si surface is
4
relatively smooth and the surface roughness of SiO after the
2
precleaning is not preserved. This film was confirmed to be
amorphous.15 On the other hand, with SiF addition, rugged
4
mountains with a larger diameter than precleaned oxide sur-
face were observed. From the combined results of TEM and
AFM, the average diameter of the rugged mountains ob-
tained from AFM does not coincide with real grain size.
Because ␮c-Si films have very small grain size from 2 to 20
nm and are mixed with both crystalline and amorphous
phases, it seems that apparent grains obtained by AFM are
coagglomerated small grains.
In summary, we have deposited the ␮c-Si films on
SiO substrates and investigated the effect of in situ hydro-
2
gen plasma precleaning. In situ hydrogen plasma is effective
in removing oxygen and carbon impurities and also induces
surface roughening which gives favorable adsorption sites
for crystalline film growth. Although SiF addition rough-
4
ened the deposited-film surface, crystalline silicon films with
͗110͘ preferential orientation could be obtained with appro-
priate amounts of SiF4.
This research has been supported by Samsung Electron-
ics Co.
FIG. 4. Three-dimensional AFM images of Si surface roughness. The scan
range is 1 ␮mϫ1 ␮m. Si H /H ϭ0.1/100, temperature at 430 °C, plasma
2
6
2
power at 60 W ͑a͒ without SiF and ͑b͒ with SiF addition at 30 sccm.
4
4
1
crystalline phase and the transition layer thickness is about
0 nm as marked with arrows in Fig. 2͑a͒. Formation of this
transition region could be suppressed with in situ plasma
cleaning as shown in Fig. 2͑c͒ with arrows. This shows that
hydrogen plasma cleaning on the initial surface strongly af-
fected the crystallization processes. The silicon film depos-
N. Shibata, K. Fukada, H. Ohtoshi, J. Hanna, S. Oda, and I. Shimizu, Jpn.
J. Appl. Phys. 26, L10 ͑1987͒.
H. Kakinuma, M. Mohri, M. Sakamoto, and T. Tsuruoka, J. Appl. Phys.
7
2
70, 7374 ͑1991͒.
3
4
5
6
7
8
9
0
1
M. Mohri, H. Kakinuma, M. Sakamoto, and H. Sawai, Jpn. J. Appl. Phys.
0, L779 ͑1991͒.
3
B. Anthony, L. Breaux, T. Hsu, S. Banerjee, and A. Tasch, J. Vac. Sci.
Technol. B 7, 621 ͑1989͒.
T. Hsu, B. Anthony, R. Qian, J. Irby, S. Banerjee, A. Tasch, S. Lin, H.
Marcus, and C. Magee, J. Electron. Mater. 20, 279 ͑1991͒.
R. J. Carter, T. P. Schneider, J. S. Montgomery, and R. J. Nemanich, J.
Electrochem. Soc. 141, 3136 ͑1994͒.
ited with SiF exhibits a columnar structure growing in a
4
direction perpendicular to the substrate. The concentric
circles of the selected area diffraction ͑SAD͒ patterns ͓Figs.
2͑b͒ and 2͑d͔͒ indicate that the silicon film deposited with
SiF is polycrystalline. By XRD and TEM results, both sili-
M. Ishii, K. Nakashima, I. Tajima, and M. Yamamoto, Appl. Phys. Lett.
4
58, 1378 ͑1991͒.
con films were found to have a preferred orientation along
the growth direction with the ͗110͘ of the film parallel to the
M. Delfino, S. Salimian, D. Hodul, A. Ellingboe, and W. Tsai, J. Appl.
Phys. 71, 1001 ͑1992͒.
T. Shibata, Y. Nanishi, and M. Fujimoto, Jpn. J. Appl. Phys. 29, L1181
͗111͘ of the substrate. The film was found to be made up of
very small needle-like grains ϳ4 nm in width and 20 nm in
length from the dark field image ͓df in Fig. 2͑c͔͒, and those
are uniformly distributed in the thickness of the film.
͑
1990͒.
1
1
Y. Kunitsuga, I. Suemune, Y. Tanaka, Y. Kan, and M. Yamanishi, J.
Cryst. Growth 95, 91 ͑1989͒.
C. W. Nam, S. Ashok, W. Tsai, and M. E. Day, J. Vac. Sci. Technol. B
Figure 3 shows AES depth profiles before and after hy-
drogen plasma cleaning. A single-crystal silicon wafer was
used because the cleaning effect can be detected more easily
on this native oxides. One of the obstacles disturbing the
growth of crystalline Si films at low temperature is the im-
purities ͑especially oxygen and carbon͒ remaining in a non-
UHV vacuum chamber which cover and react immediately
with the surface of growing silicon.16 After hydrogen plasma
cleaning oxygen as well as carbon was detected below the 1
1
2, 3010 ͑1994͒.
12
T. Nagahara, K. Fujimoto, N. Kohno, Y. Kashiwagi, and H. Kakinoki,
Jpn. J. Appl. Phys. 31, 4555 ͑1992͒.
Y. Okada, J. Chen, I. H. Campbell, P. M. Fauchet, and S. Wagner, J.
Non-Cryst. Solids 114, 816 ͑1989͒.
1
1
3
4
G. N. Parson, J. J. Boland, and J. C. Tsang, Jpn. J. Appl. Phys. 31, 1943
͑
1992͒.
15
16
X. Li, Y.-B. Park, D.-H. Kim, and S. Rhee, Mater. Lett. 24, 79 ͑1995͒.
M. K. Sanganeria, M. C. Ozturk, K. E. Violette, G. Harris, C. A. Lee, and
D. M. Maher, Appl. Phys. Lett. 66, 1255 ͑1995͒.
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:
Appl. Phys. Lett., Vol. 68, No. 16, 15 April 1996 Y.-B. Park and S.-W. Rhee 2221
130.239.20.174 On: Sat, 22 Nov 2014 11:21:01