st
21 International Symposium on Plasma Chemistry (ISPC 21)
Sunday 4 August – Friday 9 August 2013
Cairns Convention Centre, Queensland, Australia
Hydrogen radical-injection plasma fabricated
microcrystalline silicon thin film for solar cells
Masaru Hori, Yusuke Abe, Atsushi Fukushima, Ya Lu, Sho Kawashima, Keita Miwa,
Keigo Takeda, Hiroki Kondo, Kenji Ishikawa, and Makoto Sekine
1
Nagoya University, Furo-cho, Chikusa, Nagoya 464-8603 Japan
Abstract: Hydrogenated microcrystalline silicon thin films were fabricated by the plasma-enhanced chemical deposition system with hydrogen radical-injection (RICVD). The absolute density of hydrogen atom (H) was higher in the RICVD than the conventional CVD
system (C-CCP), in which measured by the vacuum ultraviolet laser absorption spectroscopy
(VUVLAS). Compared with those of C-CCP, high deposition-rates of about 2 nm/s for the
RICVD system have been realized with similar quality of crystallinity factor, preferential orientation, defect density, microstructure, and post-deposition oxidation.
Keywords: Microcrystalline silicon, Thin film solar cells, Vacuum-ultraviolet absorption
spectroscopy, Radical injection chemical vapor deposition
1. Introduction
The tandem solar cell was commoly fabricated by stack
of hydrogenated microcrystalline silicon (μc-Si:H) thin
film for the bottom cell and amorphous silicon (a-Si:H)
thin films for the top cell. The μc-Si:H can absorb light in
higher wavelength towards the infrared region of the solar
spectrum and has excellent stability against light soaking
[1–4]. However, the indirect optical transition of μc-Si:H
is not efficiently absorbed sunlight, film thicknesses
greater than 2 μm are required. To realize low-cost fabrication of solar cell with the plasma-enhanced chemical
vapor deposition (PECVD) processes at low temperature,
the high-rate growth of high-quality μc-Si:H films is required.
To date, the growth rate of approximately 2 nm/s was
realized using a capacitively coupled plasma (CCP) system with a very high frequency (VHF)-power source and
a high working pressure (ca. 1000 Pa) with a narrow electrode gap.[5–7] The high pressure and VHF-power source
provide low-ion energy and a high-electron density, which
leads to the efficient generation of a film precursor and
suppresses the formation of defects by ion bombardment.
H radicals are recognized as a key factor that influences
the crystallinity of Si thin films. Sufficient supply of H
radicals to a growing Si thin film surface induces crystallization. In particular the relatively high working pressure,
the high electron density is demanded on achievement
effectively the SiH4 depletion. Namely, only survival of
the H atom is obtained while the SiH4 depleted because
the SiH4 scavenged hydrogen atom (H) [4,8,9].
The reaction of H radicals on a growing Si film surface
was modeled as follow; the H radicals cover the growth
surface and induce local heating by recombination to hydrogen molecules [4,10]. Those processes enhance sur-
H2 plasma
SiH4/H2
60 MHz
H2
60 MHz
SiH4
MI
OES
VUVLAS
10 mm
SiH4/H2 plasma
Heater
Conventional capacitivelycoupled plasma (C-CCP)
Substrate
Heater
Radical-injection capacitivelycoupled plasma (RI-CCP)
Fig. 1 Schematic illustration of the conventional capacitively-coupled plasma (C-CCP) (left) and CCP source with
radical injection (RICVD) (right).
st
21 International Symposium on Plasma Chemistry (ISPC 21)
Sunday 4 August – Friday 9 August 2013
Cairns Convention Centre, Queensland, Australia
face diffusion of the film growth precursors. The precursors adsorbed on the sites and formed stable Si-Si bonds.
The H radicals etch the Si atoms bonded weakly to other
atom in the amorphous phase. This is resulted to reconstruct a network of Si-Si bonds in H-rich subsurface regions [11–13]. Thus when the film growth rate exceeds to
2 nm/s, the optoelectronic properties of μc-Si:H films are
degraded due to dangling-bond defects in the film
[8,14,15].
As the SiH4 highly depleted, short-lifetime species such
as SiH2, SiH, and Si contributed dominantly on the film
growth. The short-lifetime species stacked with high reactivity on the growing film surface without surface diffusion to form Si-Si bonds after insertion into Si-H bonds.
Under this circumstance, the density of dangling-bond
defects in the films increases. This means that low defect
densities in the μc-Si:H films is resulted under low depletion of SiH4. To achieve sufficiently high density of H
radical for crystallization of the thin Si films under a
low-SiH4-depletion condition, we propose to use of the H
radical-injection (RI)-PECVD for the deposition of thin Si
films.
The RI system enables the individual control of multiple species. Indeed we have demonstrated that the morphology and growth rates of the carbon nanowalls
(CNWs) has been controlled by the H radical density in
C2F6/H2 CCP using an RI-system that employs inductively coupled plasma [16,17]. The RI-PECVD system was
developed to achieve large area growth of the CNWs with
reasonable growth rates using a surface wave microwave
excited H2 plasma as a radical source [18,19]. The behavior of H radicals in ultrahigh-frequency SiH4 plasma has
been investigated using a vacuum ultraviolet absorption
spectroscopy [20].
In this study, the plasma parameters and film-growth
characteristics obtained using the PECVD system with
and without the RI system are compared. The electron
density, electron temperature, and H radical density were
measured to evaluate the features of the RI system. The
crystallinity factor, preferential orientation, defect density,
microstructure, and post-deposition oxidation of the resultant thin Si film were analyzed to evaluate and discuss
the effect of the RI system.
2. Experimental
Figure 1 shows schematic illustrations of the (a) conventional capacitively coupled plasma (C-CCP) source
and (b) CCP source with RI (RI-CCP). Both sources have
a showerhead VHF electrode, to which 60MHz VHF
power was applied, and a lower grounded (GND) electrode. A silane plasma diluted hydrogen (SiH4/H2) was
generated between these electrodes. The RI-CCP source
has an additional showerhead type upper GND electrode
above the VHF electrode. In the case of C-CCP, H2 gas
was introduced through the VHF electrode with a gas
inlet of the SiH4 gas. In the case of RI-CCP, the H2 gas
was introduced into the upper region through the upper
GND electrode to generate the H2 plasma between the
upper GND and VHF electrodes. H radicals generated in
the H2 plasma are supplied to the SiH4/H2 plasma. The
SiH4/H2 gas mixture ratio was 3% with a total gas flow
rate of 1000 sccm in both cases, and the total pressure was
kept at 1200 Pa. The distance between the VHF and lower
GND electrodes was 10 mm, and that between the VHF
Fig. 2 Schematic diagram of experimental setup for the CVD system with the vacuum-ultraviolet laser absorption
spectroscopy (VUVLAS).
st
21 International Symposium on Plasma Chemistry (ISPC 21)
Sunday 4 August – Friday 9 August 2013
Cairns Convention Centre, Queensland, Australia
and upper GND electrodes was 5 mm.
The electron density in the SiH4/H2 plasma region was
measured using a 35 GHz microwave interferometer.[21]
A multi-channel spectrometer was used to observe the
optical emission intensity of the Si* (288 nm) and SiH*
(414 nm) lines. The absolute density of H radicals was
measured by vacuum ultraviolet laser absorption spectroscopy (VUVLAS) as shown in Fig. 2. The details of
have been described in detail elsewhere [22].
Three different substrates, glass (Corning EAGLE
XGTM), quartz, and Si(100) were placed on the lower
GND electrode. The 1 μm thick Si thin films were deposited on these substrates. The substrate temperature was
373 K.
The crystallinity factor and preferential orientation of
the Si thin films deposited on glass substrates were then
analyzed by Raman spectroscopy and X-ray diffraction
(XRD). The density of defects on the films deposited on
quartz substrates was measured by electron spin resonance (ESR) spectroscopy. Post-deposition oxidation of
films deposited on Si(100) substrates was evaluated by
Fourier-transform infrared spectroscopy (FTIR). The microstructure of films was observed by transmission electron microscopy (TEM).
3. Results and discussion
For RI-CCP, both the SiH4/H2 and H2 plasmas are produced with one VHF power; therefore, the electron density should be different from that for C-CCP at the same
VHF power. The electron density as a function of the
VHF power in the SiH4/H2 plasma region was measured
for comparison of RI-CCP and C-CCP under the same
plasma conditions. As the results, the electron density
increased from 4.7 x 1010 to 12.8 x 1010 cm-3 for C-CCP
and from 2.0 x 1010 to 9:3 x 1010 cm-3 for RI-CCP with an
increase in the VHF power from 200 to 700 W. The densities for RI-CCP are lower (44–72%) than those for C-CCP
at the same VHF power. Thus comparisons between
RI-CCP and C-CCP at the same electron density were
carried out.
H radicals play an important role in the formation of Si
thin films. The absolute density of H radical in the
SiH4/H2 plasma was measured using the VUVLAS technique. The density of H radicals increased from 4.8 x 1012
to 8.2 x 1012 cm-3 for C-CCP and from 5.5 x 1012 to 8.3 x
1012 cm-3 for RI-CCP with an increase in the electron
density. The density of H radicals with RI-CCP is ca. 1.4
times higher than that with C-CCP at the same electron
density, due to the injection of H radicals supplied from
the H2 plasma with RI-CCP. The major transport mechanism of H radicals is the diffusion. The flux of H radicals
by the diffusion is four orders of magnitude higher than
that by gas flow. The RI system enables an increase in the
H radical density at the same electron density and temperature.
The deposition rate for C-CCP increased slightly from
2.8 to 3.2 nm/s with increasing electron density, and then
became saturated. The deposition rate for RI-CCP increased from 2.0 to 3.0 nm/s with increasing electron
density and became saturated. The deposition rates satisfies the requirement in the production-level for the fabrication of solar cell devices. This resulted due to enhance
dissociation of SiH4 as the electron density increased.
Saturation of the deposition rate indicates that the silane
species is depleted. The deposition rates for both plasmas
were comparative at the same electron density. Only the
production of the film growth precursors is strongly dependent on the plasma density. Oppositely, the etching
effect of H radicals is ignorable enough low of etch yield
of 0.01 [23], and H radical flux of 1017 cm-2s-1 calculated
from the results of VUVLAS measurement. The deposition rates of more than 1 nm/s were one order higher in
magnitude than etching rate of 0.1 nm/s here. This indicates that the H etching has no significant contribution on
deposition even though the H radical density was higher
of 1.4 times for the RI-CCP than that for C-CCP.
The mechanism to determine the crystalline orientation
has not completely understood yet. The composition of
deposition radicals (SiH3 and polysilane radicals) may
play an important role in the determination of the growth
facets [24,25] For RI-CCP, crystallization occurs under
low depletion of SiH4. Therefore, the ratio of SiH3 radicals to polysilane radicals may be different from that for
C-CCP.
The degree of <110> preferential orientation,
I(220)=I(111), where I(hkl) denotes the integrated intensity of the (hkl) phase in the XRD pattern was evaluated.
The Si thin film grown at an electron density of 6.3 x 1010
cm-3 by RI-CCP has weak <110> preferential orientation.
It is necessary to measure deposition radicals to investigate the mechanism of <110> preferential orientation. As
H radical density changes in the RI-system as shown before, film structure in crystallinity factor and orientation
at the transition region between a-Si:H and lc- Si:H were
apparently modified. This is due to that the film structure
saturated to change under conditions at the high H radical
flux [26]. μc-Si:H films with low defect densities (5.1 x
1016 cm-3) and low post-deposition oxidation can be
achieved using RI-CCP, due to less short-lifetime species
and less ion bombardment.
Crystallization was realized under a low electron density or low depletion of SiH4 with RI-CCP, due to a sufficient H radicals supplied by the RI system. Thus, RI-CCP
has a high possibility to realize high-quality μc-Si:H films
for the fabrication of solar cell devices.
4. Conclusion
We proposed the H Radical-injection PECVD system
excited at 60MHz for the fabrication of Si thin films for
solar cell devices to achieve selective enhancement of the
st
21 International Symposium on Plasma Chemistry (ISPC 21)
Sunday 4 August – Friday 9 August 2013
Cairns Convention Centre, Queensland, Australia
H radical densities for crystallization of the films under
low depletion of SiH4. The plasma parameters and film
growth characteristics of RI-CCP were compared with
C-CCP. The electron density was measured using a microwave interferometer. The absolute density of H radicals was measured by VUVLAS. The crystallinity factor,
preferential orientation, defect density, microstructure,
and the post-deposition oxidation of deposited Si thin
films were investigated by Raman spectroscopy, XRD,
ESR, TEM, and FTIR, respectively. From these results,
the following features of the RI system were determined:
1. The density of H radicals in RI-CCP was ca. 1.4
times higher than that in C-CCP at the same electron density (5.5 x 1012 – 8.3 x 1012 cm-3).
2. The crystallinity factor of 0.6 was achieved under a
low electron density (4.1 x 1010 cm-3) or a low depletion
of SiH4.
3. Si thin film deposited with RI-CCP showed weak
<110> preferential orientation (I(220)=I(111)=0.9).
4. The defect density by crystallization with RI-CCP is
lower than that with C-CCP (5.1 x 1016 cm-3).
The high performance of RI-CCP for the deposition of
Si thin films for solar cell devices has been demonstrated.
The RI-system was constructed with one power supply of
60MHz, which is divided into two regions, so that equipment costs and size could be kept down. This equipment
is a novel system for the fabrication of μc-Si:H films for
solar cell devices. Furthermore, the results obtained in this
study should be useful and give the guideline from viewpoints of basic and applied research for the development
of high quality films.
5. References
[1] J. Meier, R. Fluckiger, H. Keppner, and A. Shah,
Appl. Phys. Lett. 65, 860 (1994).
[2] H. Keppner, J. Meier, P. Torres, D. Fischer, and A.
Shah, Appl. Phys. A 69, 169 (1999).
[3] K. Yamamoto, T. Suzuki, M. Yoshimi, and A.
Nakajima, Jpn. J. Appl. Phys. 36, L569 (1997).
[4] A. Matsuda, J. Non-Cryst. Solids 338–340, 1 (2004).
[5] L. Guo, M. Kondo, M. Fukawa, K. Saitoh, and A.
Matsuda, Jpn. J. Appl. Phys. 37, L1116 (1998).
[6] M. Kondo, M. Fukawa, L. Guo, and A. Matsuda, J.
Non-Cryst. Solid. 266–269, 84 (2000).
[7] U. Graf, J. Meier, U. Kroll, J. Bailat, C. Droz, E.
Vallat-Sauvain, and A. Shah, Thin Solid Films 427,
37 (2003).
[8] M. Fukawa, S. Suzuki, L. Guo, M. Kondo, and A.
Matsuda, Sol. Energy Mater. Sol. Cells 66, 217
(2001).
[9] A. Gordijn, M. Vanecek, W. J. Goedheer, J. K. Rath,
and R. E. I. Schropp, Jpn. J. Appl. Phys. 45, 6166
(2006).
[10] Y. Toyoshima, K. Arai, A. Matsuda, and K. Tanaka,
Appl. Phys. Lett. 56, 1540 (1990).
[11] C. C. Tsai, G. B. Anderson, R. Thompson, and B.
Wacker, J. Non-Cryst. Solids 114, 151 (1989).
[12] F. Kail, A. F. I. Morral, A. Hadjadj, P. R. I.
Cabarrocas, and A. Beorchia, Philos. Mag. 84, 595
(2004).
[13] K. Nakamura, K. Yoshino, S. Takeoka, and I.
Shimizu, Jpn. J. Appl. Phys. 34, 442 (1995).
[14] Y. Sobajima, S. Nakano, T. Toyama, and H. Okamoto,
Jpn. J. Appl. Phys. 46, L199 (2007).
[15] M. Kondo, Sol. Energy Mater. Sol. Cells 78, 543
(2003).
[16] M. Hiramatsu, K. Shiji, H. Amano, and M. Hori,
Appl. Phys. Lett. 84, 4708 (2004).
[17] K. Shiji, M. Hiramatsu, A. Enomoto, M. Nakamura,
H. Amano, and M. Hori, Diamond Relat. Mater. 14,
831 (2005).
[18] S. Kondo, M. Hori, K. Yamakawa, S. Den, H. Kano,
and M. Hiramatsu, J. Vac. Sci. Technol. B 26, 1294
(2008).
[19] W. Takeuchi, M. Ura, M. Hiramatsu, Y. Tokuda, H.
Kano, and M. Hori, Appl. Phys. Lett. 92, 213103
(2008).
[20] S. Takashima, M. Hori, T. Goto, and K. Yoneda, J.
Appl. Phys. 89, 4727 (2001).
[21] G. Neumann, U. Banziger, M. Kammeyer, and M.
Lange, Rev. Sci. Instrum. 64, 19 (1993).
[22] Y. Abe, S. Kawashima, K. Takeda, M. Sekine, and M.
Hori, Appl. Phys. Express 3, 106001 (2010).
[23] G. Dingemans, M. N. van den Donker, D. Hrunski, A.
Gordijn, W. M. M. Kessels, and M. C. M. van de
Sanden, Appl. Phys. Lett. 93, 111914 (2008).
[24] K. Saito and M. Kondo, Phys. Status Solidi A 207,
535 (2010).
[25] H. Kakinuma, M. Mohri, M. Sakamoto, and T.
Tsuruoka, J. Appl. Phys. 70, 7374 (1991).
[26] S. Nunomura and M. Kondo, J. Phys. D 42, 185210
(2009).