Combinatorial study of substrate temperature dependence on properties
of silicon films deposited using multi-hollow discharge plasma CVD
Yeonwon Kim, Takeaki Matsunaga, Yuki Kawashima, Daisuke Yamashita, Kunihiro Kamataki, Naho Itagaki,
Giichiro Uchida, Kazunori Koga, and Masaharu Shiratani
Graduate School of Information Science and Electrical Engineering, Kyushu University, Japan
[email protected]
Abstract: A combinatorial technique using a multi-hollow discharge plasma CVD method has been
employed to investigate effects of substrate temperature on hydrogenated microcrystalline silicon (cSi:H) film growth. No film is deposited near the discharge region RT to 350°C, whereas the area of no
film region decreases with increasing the substrate temperature. This is attributed to competitive
reactions between Si etching by H atoms and Si deposition. The area of microcrystalline Si region
increases with the substrate temperature due to high surface diffusion rate of Si containing radical. These
results show that the process window of μc-Si:H films becomes wider for the higher substrate
temperature due to two reasons: 1) the lower Si etching rate and 2) the longer surface migration length of
Si containing radicals. The defect density decreases significantly with increasing Ts from 6.24 × 1016
cm-3 for Ts =100oC down to 6.26 × 1015 cm-3 for Ts =300oC.
Keywords: combinatorial study, multi-hollow CVD method microcrystalline silicon, temperature dependence
Tandem Si thin film solar cells are
fabricated using a-Si:H cells as the top cells and
μc-Si:H as the bottom cells, in order to realize
high efficiency with low manufacturing cost [1].
High rate deposition of hydrogenated
microcrystalline silicon (μc-Si:H) film has been
the subject of intensive studies from both
fundamental, experimental, and applied interests
[2-4]. Recently, high pressure depletion regime
using plasma CVD has been proposed as a
deposition method leading to much higher
growth rates while preserving acceptable film
quality [5]. Optimization of μc-Si:H solar cells
requires deeper understanding of the material
properties in the high depletion regime,
especially the process window [6] of high
quality microcrystalline films determined by
etching by H [7, 8] and surface diffusion of
deposition radicals [9].
Here we report experimental results of
combinatorial study of substrate temperature
dependence on film properties using a multihollow discharge plasma CVD method.
2. Experimental details
Figure 1 shows a schematic diagram of the
multi-hollow discharge plasma CVD reactor [10].
Multi-hollow electrodes had 24 holes of 5mm in
diameter, in which the discharges were sustained.
VHF power of 180W (60MHz) was applied between
one powered electrode and a pair of grounded
electrodes to generate plasma. SiH4 diluted with H2
was supplied to the reactor. The silane dilution ratio
R= [SiH4]/([H2]+[SiH4]) were kept at 0.997. The
total pressure was 2 Torr. The substrate holder was
set vertically on the powered electrode to obtain
information on spatial distribution of deposition
distance from electrode z (mm)
1. Introduction
pump
SiH3
film
quartz substrate
grounded substrate holder
H
discharge region
powered electrode
grounded electrode
50
40
30
20
10
0
～
gas inlet
Figure 1 Schematic diagram of multi-hollow discharge plasma
CVD reactor.
RT
o
100 C
o
250 C
o
200 C
o
300 C
o
350 C
0.1
0.01
deposition rate (nm/s)
(a)
(a)
0.1
0.01
(b)
(b)
0.8
width(mm)
40
c
crystalline volume
fraction (X )
deposition rate (nm/s)
Figure 2 Images of films prepared by the multi-hollow discharge plasma CVD method as a parameter of substrate temperature, Ts=RT350oC, 180W (60MHz),R=0.997, 2Torr.(White dotted lines indicate the boundary between crystallized and amorphous region)
0.6
0.4
crystallized region
suitable X region (X =0.5 - 0.7)
c
c
30
20
10
0.2
0
0
0
5
10 15 20 25 30 35 40
distance from electrode z (mm)
Figure 3 Deposition rate (a) and crystalline volume fraction
(Xc) (b) as a function of distance from electrode.
radicals, etching precursors and film properties. The
substrate temperatures (Ts) were varied from RT to
350oC. Crystalline volume fraction and crystalline
orientation were determined with a Raman
spectroscope (Jasco, NRS-3100) and X-ray
diffraction (XRD) spectrometer (Bruker, D8
DISCOVER), respectively. Deposition rate was
obtained as film thickness divided by deposition
time. Defect density was measured with an electron
spin resonance (ESR) spectrometer (Bruker, EMX)
[11, 12].
3. Results and discussion
Figure 2 shows images of films as a parameter of
substrate temperature. No film is deposited near the
discharge region in all temperature range and
0
50 100 150 200 250 300 350 400
substrate temperature( oC)
Figure 4 Substrate temperature (Ts) dependence of deposition
rate (a) and width of crystallized region (b).
whereas the area of no film region decreases with
increasing Ts. This is attributed to competitive
reactions between Si etching by H atoms and Si
deposition [13]. The deposition rate decreases
exponentially with z for z>5 cm [Fig. 3 (a)], because
SiH3 density [14-16], which is the main deposition
precursor for high-quality films, decreases
exponentially with z due to their loss at the surfaces.
The crystalline volume fraction (XC) was deduced
from the integrated intensity ratio, XC =(I520+I500)/
(I520+I500+I480), by de-convolution of the Raman
spectra into three characteristic Raman peaks
positioned at 480, 500 and 520 cm-1[17]. In Fig. 3(b)
we can clearly identify the crystallized region and
the amorphous region. White dotted lines in Fig. 2
show the boundary between these two regions. The
crystalline volume fraction is high near the discharge
17
1x10
5
defect density (cm )
(a)
I(220)/I(111)
-3
4
3
2
1
16
1x10
crystallite size (Å)
0
(b)
100
100 150 200 250 300 350
400
o
Substrate temperature ( C)
Figure 6 Substrate temperature (Ts) dependence of ccrystalline film (Xc=0.5-0.7).
50
0
50
0
50 100 150 200 250 300 350 400
substrate tempertature( oC)
Figure 5 Substrate temperature (Ts) dependence of crystal
orientation ratio of (220) to (111), I220/I111,(a) and (220)
crystallite size (b).
region and drops sharply in a transition region
around z=20-40mm [Fig. 3(b)]. These results
suggest a flux ratio of H to SiH3 radicals is high near
the discharge region and decreases with z, because a
high flux ratio is needed for c-Si:H film deposition
[3]. The area of microcrystalline Si region increases
with Ts due to high surface diffusion rate of
deposition radicals [18]. These results show that the
process window of μc-Si:H films becomes wider for
the higher substrate temperature due to two reasons:
1) the lower Si etching rate and 2) the longer surface
migration length of Si containing radicals.
Figure 4(a) shows that deposition rate of
Xc=0.5-0.7 region as a function of Ts. It increases
with Ts up to 200oC, nearly constant for 200-300oC,
and decreases up to Ts of 350oC. Figure 4(b) shows
the width of crystallized region (Xc>0) and that of
Xc=0.5-0.7, crystalline volume fraction profiles in
Fig. 3(b). They increase up to Ts of 200oC, then
decrease to Ts of 300oC.
To obtain more information on the film structure,
we evaluated crystalline orientation from XRD
spectra using the peak intensity ratio of the (220)
orientated c-Si:H to the (110) orientated one,
I220/I111. The (220) orientated crystallite size was
deduced from the XRD peak width, using the
Scherrer’s equation [19]. Figure 5(a) shows the Ts
dependence of the film orientation in Xc=0.5-0.7
film region. The (111) orientated c-Si:H films are
obtained preferentially at relatively low Ts of 100oC,
and that the preferential crystal orientation changes
from (111) to (220) at Ts of 200oC and maintained
the constant value over Ts of 200oC, The (200)
crystallite size increases slightly with Ts[Fig. 5 (b)].
Figure 6 shows Ts dependence of defect density
Ns of films of the Xc=0.5-0.7 region. The defect
density has a high value of 6.24  1016 cm-3 for Ts
=100oC, it decreases significantly with Ts down to
6.26  1015 cm-3 for Ts 300oC and then slightly
increased up to Ts=350oC.
4. Conclusions
The combinatorial technique using the multihollow discharge plasma CVD method have been
used to investigate the effects of substrate
temperature on the c-Si:H film growth. No film is
deposited near the discharge region at low substrate
temperature from RT to 100°C, whereas the area of
no film region decreases with increasing the
substrate temperature. This is attributed to
competitive reactions between Si etching by H
atoms and Si deposition. The area of
microcrystalline Si region increases with the
substrate temperature due to high surface diffusion
rate. These results show that the process window of
μc-Si:H films becomes wider for the higher substrate
temperature due to two reasons: 1) the lower Si
etching rate and 2) the longer surface migration
length of Si containing radicals. The defect density
has a high value of 6.24 × 1016 cm-3 for Ts =100oC,
it decreases significantly with Ts down to 6.26 × 1015
cm-3 for Ts=300oC.
[16] O. Vetterl, F. Finger, R. Carius, P. Hapke, L. Houben,
O. Kluth, A. Lambertz, A. Mück, B. Rech, and H.
Wagner, Sol. Energy Mater. Sol. Cells 62, 97(2003).
Acknowledgements
[17] L. Houben, M. Luysberg, P. Hapke, R. Carius, F.
Finger and H. Wagner, Philos. Mag. A 77,
1447(1998).
This work was partly supported by the Industrial
Technology Development Organization (NED), and
MEXT
References
[1] J. Meir, S. Dubail, R. Fluckiger, H. Keppner, and A.
Shah, Proc. 1st WCPWC, 409 (1994).
[2] Y. Sobajima, T. Higuchi, J. Chantana, T. Toyama, C.
Sada, A. Matsuda, and H. Okamoto, Phys. Status
Solidi C 7, 3–4, 521–524 (2010)
[3] S. Nunomura, I. Yoshida, M. Kondo, Appl. Phys. Lett.
94, 071502 (2009).
[4] Y. Sobajima M. Nishino, T. Fukumori, T. Higuchi, S.
Nakano, T. Toyama, H. Okamoto, Sol. Energy Mater.
Sol. Cell 93, 980 (2009).
[5] M. Kondo, M. Fukawa, L. Guo, A. Matsuda, J. NonCryst. Solids 84, 266-269(2000).
[6] S. Klein, F. Finger, R. Carius, M. Stutzmann, J. Appl.
Phys. 98, 024905(2005).
[7] X. R. Duan, H. Lange, A. Meyer-Plath, Plasma
Sources Sci. Technol.12, 554(2003).
[8] P. Kae-Nune, J. Perrin, J. Jolly, and J. Guillon, Surf.
Sci. 360, L495(1996).
[9] W. Beyer, Sol. Energy Mater. Sol. Cells 78, 235-67
(2003).
[10] K. Koga, T. Inoue, K. Bando, S. Iwashita, M.
Shiratani, and Y. Watanabe, Jpn. J. Appl. Phys. 44,
L1430 (2005)
[11] M. H. Brodsky, R.S. Title, Phys. Rev. Lett. 23,
581(1969).
[12] S. T. Pantelides, Phys. Rev. Lett. 57, 2979(1986).
[13] J. Abrefah, D.R. Olander, Surf. Science. 209, 291313(1989).
[14] Itanashi, N. Nishikawa, M. Magane, S. Naito, T.
Goto, A. Matsuda, C. Yamada, and E. Hirota, Jpn. J.
Appl. Phys. 29, L505-L507(1990).
[15] J. Perrin, M. Shiratani, P. Kae-Nune, H. Videlot, and
J. Guillon, J. Vac. Sci. Technol. A, Vac. Surf. Films
16, 278-289(1998).
[18] P. C. P. Bronsveld, J. K. Rath, and R. E. I. Schropp,
et al., Appl. Phys. Lett. 89, 051922 (2006).
[19] S. Vepřek, F.-A. Sarott and M. Rückschloß, J. NonCryst. Solids 137&138, 733-736(1991).