Crystal Structure Theory and Applications, 2012, 1, 57-61 Published Online December 2012 (
Amorphous-Nanocrystalline Transition in Silicon Thin
Films Obtained by Argon Diluted Silane PECVD
Rachid Amrani1,2*, Frederic Pichot1, Larbi Chahed2, Yvan Cuminal1
1Université Montpellier II, Place Eugène Bataillon Montpellier, Montpellier, France
2Département de Physique, Université d’Oran ES-Sénia, Oran, Algérie
Email: *,
Received October 4, 2012; revised November 8, 2012; accepted November 22, 2012
The Plasma-Enhanced Chemical Vapor Deposition (PECVD) method is widely used compared to other methods to de-
posit hydrogenated silicon Si:H. In this work, a systematic variation of deposition parameters was done to study the
sensitivities and the effects of these parameters on the intrinsic layer material properties. Samples were deposited with
13.56 MHZ PECVD through decomposition of silane diluted with argon. Undoped samples depositions were made in
this experiment in order to obtain the transition from the amorphous to nanocrystalline phase materials. The substrate
temperature was fixed at 200˚C. The influence of depositions parameters on the optical proprieties of the thin films was
studied by UV-Vis-NIR spectroscopy. The structural evolution was also studied by Raman spectroscopy and X-ray dif-
fraction (XRD). The structural evolution studies show that beyond 200 W radio frequency power value, we observed an
amorphous-nanocrystalline transition, with an increase in crystalline fraction by increasing RF power and working
pressure. The deposition rates are found in the range 6 - 10 Å/s. A correlation between structural and optical properties
has been found and discussed.
Keywords: Silicon Thin Film; PECVD; Amorphous-Nanocrystalline Transition; Deposition Rate; Powders; Argon
1. Introduction
The hydrogenated nanocrystalline silicon (nc-Si:H) has
gained much attention over amorphous silicon (a-Si:H).
The light-induced degradation of a-Si:H, also named Sta-
bler-Wronski effect [1,2], has been a major bottleneck
since it was observed in practical device applications
such as thin film solar cells. Several deposition methods
have been used to prepare device quality films. These in-
clude Hot Wire-CVD (chemical vapor deposition) [3],
Electron Cyclotron Resonance-CVD [4], conventional
Plasma-Enhanced-CVD (PECVD) [5], Very High Fre-
quency Plasma Enhanced-CVD [6], Microwave-CVD [7]
and Radio-Frequency Magnetron Sputtering of crystal-
line silicon target [8]. Among these, only PECVD has
been established for industrial applications.
However, the device quality nc-Si:H films prepared by
PECVD method with silane hydrogen mixture, at opti-
mized deposition parameters show lower deposition rate
(0.4 - 3 Å/s) .
Another problem with this kind of method is the pow-
ders formation during the film growth. A general study
of powders formation has been performed by Dorier et al.
[9-11] as a function of silane dilution in argon, helium or
hydrogen in PECVD system. Dorier [12] compared the
powder appearance time value ta versus gas used for si-
lane dilution; argon, helium or hydrogen. As shown in Fig-
ure 1, it is interesting to note that a strong difference is
evidenced when comparing between these three situa-
tions. ta decreases when silane is diluted in argon while
helium and hydrogen dilution lead to the inverse effect.
The probability that initial clusters have a chance to form
small particles being stored in the plasma is enhanced for
argon dilution and reduced for helium or hydrogen dilu-
Figure 1. Particle appearance time as a function of silane
partial pressure in silane mixtures with argon, helium or
hydrogen [12].
*Corresponding author.
opyright © 2012 SciRes. CSTA
When silane is diluted with hydrogen, it is well known
that growth rate is reduced, with simultaneous trends to
form nanocrystalline structure in place of a-Si:H. The
lower deposition rates increase the process operation
time and hence the production cost.
In this work, the Si:H thin films were deposited in a
conventional PECVD system, using argon silane mixture.
By Raman and XRD spectroscopy, the structural evolu-
tion was studied. The optical characterization for these
films was also appraised by UV-Visible-NIR spectropho-
2. Experimental Procedure
Samples were deposited in a Radio-Frequency (RF)
13.56 (MHZ) PECVD chamber. The electrode area was
350 cm2 and the distance between electrodes 1.2 cm. The
mixed precursor gases were ultrahigh purity silane (SiH4)
and argon (Ar). The dilution ratio (the argon to silane
flow ratio) was fixed at 25 during all the processes. The
deposition pressure varied from 400 mTorr to 1400
mTorr, at various RF powers (50, 100, 200 and 250 W)
and substrate temperature was fixed at 200˚C.
Before the deposition, all glass substrates were cleaned
strictly following a standard procedure to ensure a good
adhesion between the substrate and the Si:H layer. Ra-
man spectra were recorded with micro-Raman spectros-
copy (Jobin Yvon Horibra LABRAM-HR). The spec-
trometer has backscattering geometry for detection of
Raman spectrum with the resolution of 1 cm–1. The exci-
tation source was 473.5 nm line of Argon laser. The
power of the Raman laser was kept about 2 mW to avoid
laser induced crystallization on the films. Low angle
X-ray diffraction pattern was obtained by X-ray dif-
fractometer using CuKα line (λ = 1.54056 Å). Thickness
and refractive index were determined by UV-Vis-NIR
spectroscopy using the envelope method [13,14]. The
average band gap EM was estimated using the procedure
followed by Wemple et al. [15].
3. Results and discussion
The variation of deposition rate plotted as a function of
process pressure is shown in Figure 2 . It shows that, the
increase of RF power increases the deposition rate. It is
seen also that the deposition rate increases from ~ 6 Å/s
to ~7.5 Å/s, when the process pressure increases from
400 mTorr to 1000 mTorr. With further increase in pro-
cess pressure to 1400 mTorr, the deposition rate de-
creases. Thus, with increase in process pressure the im-
pingement rate of silane on the plasma increases. As a
result, the number of film-forming radicals and hence the
deposition rate increases. With further increase in pro-
cess pressure the supply of film-forming radicals also
increases. However, the powders electrostatically trapped
in the plasma prevent the films from growth. So, in order
to obtain a deposition without dust trapped in the plasma,
the process pressure must be below 1000 mTorr.
The crystallinity of the films was studied by low angle
X-ray diffraction (XRD). XRD patterns of two typical
samples at the same pressure (1400 mTorr) with different
RF power are shown in Figure 3. At 100 W RF power,
no crystal grains are detected in this sample; there is no
apparent structural evolution in the thin films.
When RF power reaches 200 W, diffraction peaks
arise. With the increase of RF power, there appear three
peaks symbolizing three different silicon crystalline ori-
entations. Also, the growth of grains in the thin films is
multi-oriented. The average crystallite size (dXRD) can be
estimated using the classical Scherrer’s formula [16].
The estimated average crystallite size dXRD obtained for
the films deposited at RF power 200 W and at working
pressure of 800, 1000, 1200 and 1400 mTorr are 7.1, 9.5,
12.2 and 15.1 nm, respectively.
Micro-Raman spectroscopy has been widely used as a
powerful technique to characterize deposited thin layers.
It can elucidate the amorphous and crystalline phase.
Figure 4 shows Raman spectra of Si:H films deposited at
various process pressure and RF power.
Figure 2. Variation of deposition rate as a function of pro-
cess pressure and RF power for Si:H films deposited by
Figure 3. Low angle X-ray diffraction pattern of two typical
samples deposited at the same pressure (1400 mTorr) with
different RF power.
Copyright © 2012 SciRes. CSTA
For samples deposited below 200 W RF power, the
TO band is composed by only a broad peak located
around 480 cm–1, characteristic of a completely amor-
phous structure. Beyond 200 W RF power and beyond
800 mTorr pressure deposition, the TO band can be cor-
rectly fitted using three Gaussian components centered
around 480, 510 and 520 cm–1, suggesting the presence
in these films a mixture of amorphous as well as crystal-
line structure with different grain sizes. The estimated
crystalline volume fraction Fc [17,18] and crystallite size
dRaman [19] in the films are also indicated in Figure 4. For
nanocrystalline materials, with increase process pressure
and RF power, both, Fc and dRaman in the films increase.
For samples deposited at 200 W, Fc increases from 21%
to 90% and dRaman increases from 6.4 nm to 16.2 nm,
when the process pressure increases from 800 mTorr to
1400 mTorr.
Film particle sizes measured by XRD method turned
out significant difference with those measured by Raman
method. The difference can be due to the different detec-
tion sensitivity of characterization techniques. However,
it is important to note that the crystallite sizes determined
by both techniques at various process pressure show
same trend. Therefore, for nanocrystalline samples, with
increase in process pressure, the grain size increases.
To achieve a better understanding of the optical pro-
prieties of Si:H, the optical transmission of films was
Figure 4. Typical Raman spectra obtained in the TO-like
mode, for films deposited (a): with deposition pressure of
1000 mTorr and varying RF power (50, 100, 200 and 250
W); and (b): RF power 200 W and varying pressure (600,
800, 1000 mTorr).
measured by UV-Vis-NIR spectrophotometer. The films
thickness and the refraction indexes were determined
using the method proposed by Swanepoel [13].
Figure 5 shows the variation of static refractive index
as a function of deposition parameters. The refractive
index increases when the working pressure increases.
Also, the refractive index decreases when RF power in-
The refraction index is an important parameter, since it
could be linked to the material density [8,20]. Indeed, a
good material is compact, i.e. it presents a minimum of
Detailed analysis of the refractive index spectra were
performed using the model suggested by Wemple and
Didomenico [15]. At energies below than of the optical
bandgap, the refractive index is related to the square of
the photon energy (ħω)2 by:
 
The plot of 1/[n2(ħω) 1] versus (ħω)2 allows the de-
termination of the average gap EM, and the dispersion
energy ED.
Figure 6 shows that all samples deposited at 100 W
RF power have a high average gap value EM, which exhi-
bits a completely amorphous structure. But for films grown at
Figure 5. Variation of static refractive index for samples
deposited at different process pressure and RF power.
Figure 6. Variation of average band gap for Si:H films de-
posited by PECVD as a function of process pr essure and RF
Copyright © 2012 SciRes. CSTA
200 W and beyond 800 mTorr, EM decreases considera-
bly. This confirms once again, that beyond these deposi-
tion parameters, an amorphous-nanocrystalline transition
is observed.
The RF bias voltage is an important deposition pa-
rameter of nc-Si:H films. During the growth, the RF bias
has a decisive effect on the energy and natures of (SiHn)
species [19]. The appropriate bias may control the reac-
tion rate at the gas-substrate interface, so as to promote
the nucleation rate, and enhance the densification of
nanocrystallites in the growing films. He et al. demon-
strate also, that bias voltage controls the crystalline frac-
tion for nc-Si:H samples [19]. In this work, and at certain
deposition parameters, the bias varies. For power more
than 200 W and pressure below 800 mTorr, RF bias in-
creases from 3 V to 80 V, after a few minutes for the
beginning of the deposition process. The reflect RF
power is also an important parameter. After a specific
deposition time (~10 min), the reflect power increases,
and in consequence the real power deposition decreases.
To resolve the effects of RF bias and RF power reflect
values, samples were deposited with layer by layer (LBL)
deposition method.
4. Conclusion
As conclusion, we have proved that the formation of
powders in Silane-Argon plasma, initially considered as
a drawback, can be exploited to produce nanocrystalline
silicon particles. We have shown that hydrogenated
nanocrystalline silicon (nc-Si:H) films can be prepared
from silane argon mixture without hydrogen dilution at
high deposition rates (~10 Å/s) and at low substrate tem-
perature (200˚C) using PECVD method. The structural
evolution studies show that beyond 200 W RF power and
beyond 800 mTorr pressure deposition conditions, we
observed an amorphous-nanocrystalline transition. From
the present study it has been concluded that the process
pressure and RF power are the key process parameters to
induce the crystallinity in the Si:H films grown by
PECVD method. From materials and devices application,
especially, photovoltaic devices application, nanocrystal-
line silicon produced at low temperature in silane plasma,
open the way to new applications, like the deposition on
flexible substrate.
5. Acknowledgements
The authors would like to thank D. Bourgogne for Ra-
man experiments assistance and B. Fraisse for the X-ray
diffraction measurements. This work is kindly supported
by Averroes Program.
[1] W. E. Spear and P. G. LeComber, “Subtitutional Doping
of Amorphous Silicon,” Solid State Communications, Vol.
17, No. 9, 1975, pp. 1193-1196.
[2] D. L. Staebler and C. R. Wronki, “Reversible Conduc-
tivity Changes in Discharge Produced Amorphous Si,”
Applied Physical Letters, Vol. 31, No. 4, 1977, p. 292.
[3] H. Li, R. H. Franken, R. L. Stolk, C. H. M. van der Werf,
J. K. Rath and R. E. I. Schropp, “Controlling the Quality
of Nanocrystalline Silicon Made by Hot-Wire Chemical
Vapor Deposition by Using a Reverse H2 Profiling Tech-
nique,” Journal of Non-Crystalline Solids, Vol. 354, No.
19-25, 2008, pp. 2087-2091.
[4] M. Birkholz, B. Selle, E. Conrad, K. Lips and W. Fuhs,
“Evolution of Structure in Thin Microcrystalline Silicon
Films Grown by Electron-Cyclotron Resonance Chemical
Vapor Deposition,” Journal of Applied Physics, Vol. 88,
No. 7, 2000, pp. 4376-4379. doi:10.1063/1.1289783
[5] B. Rech, T. Roschek, J. Müller, S. Wieder and H. Wagner,
“Amorphous and Microcrystalline Silicon Solar Cells
Prepared at High Deposition Rates Using RF (13.56 MHz)
Plasma Excitation Frequencies,” Solar Energy Materials
and Solar Cells, Vol. 66, No. 1-4, 2001, pp. 267-273.
[6] Y. Mai, S. Klein, R. Carius, L. Houben, X. Geng and F.
Finger, “Improvement of Open Circuit Voltage in Micro-
crystalline Silicon Solar Cells Using Hot Wire Buffer
Layers,” Journal of Non-Crystalline Solids, Vol. 352, No.
9-20, 2006, pp. 1859-1862.
[7] M. van Veen, C. van der Werf and R. Schropp, “Tandem
Solar Cells Deposited Using Hot-Wire Chemical Vapor
Deposition,” Journal of Non-Crystalline Solids, Vol. 338-
340, No. 1, 2004, pp. 655-658.
[8] R. Amrani, D. Benlekehal, R. Baghdad, D. Senouci, A.
Zeinert, K. Zellama, L. Chahed, J. D. Sib and Y. Bouizem,
“Low-Temperature Growth of Nanocrystalline Silicon
Films Prepared by RF Magnetron Sputtering: Structural
and Optical Studies,” Journal of Non-Crystalline Solids,
Vol. 354, No. 19-25, 2008, pp. 2291-2295.
[9] J. L. Dorier, C. Hollenstein, A. A. Howling and U. Kroll,
“Powder Dynamics in Very High Frequency Silane Plas-
mas,” Journal of Vacuum Science & Technology, Vol.
A10, No. 4, 1992, pp. 1048-1052. doi:10.1116/1.578200
[10] J. L. Dorier, C. Hollenstein and A. A. Howling, “Spatio-
temporal Powder Formation and Trapping in Radio Fre-
quency Silane Plasmas Using Two-Dimensional Polari-
zation-Sensitive Laser Scattering,” Journal of Vacuum
Science & Technology, Vol. A13, No. 3, 1995, pp. 918-
926. doi:10.1116/1.579852
[11] A. A. Howling, J. L. Dorier, C. Hollenstein, U. Kroll and
F. Finger, “Frequency Effects in Silane Plasmas for
Plasma Enhanced Chemical Vapor Deposition,” Journal
of Vacuum Science & Technology, A10, No. 4, 1992, pp.
1080-1085. doi:10.1116/1.578205
[12] J. L. Dorier,“Genèse, Croissance et Conséquences de
Particules Dans les Plasmas en Silane à Basse Pression et
Basse Température,” Ph.D. Thesis, E.P.F.L, Switzerland,
Copyright © 2012 SciRes. CSTA
Copyright © 2012 SciRes. CSTA
[13] R. Swanpoel, “Determination of the Thickness and Opti-
cal Constants of Amorphous Silicon,” Journal of Physics
E: Scientific Instruments, Vol. 16, No. 12, 1983, pp. 1214-
1222. doi:10.1088/0022-3735/16/12/023
[14] J. C. Manifacier, J. Gasiot and J. P. Fillard, “A Simple
Method for the Determination of the Optical Constants n,
k and the Thickness of a Weakly Absorbing Thin Film,”
Journal of Physics E: Scientific Instruments, Vol. 9, No.
11, 1976, pp. 1002-1004.
[15] S. H. Wemple and M. Didomenico, “Behavior of the
Electronic Dielectric Constant in Covalent and Ionic Ma-
terials,” Physical Review B, Vol. 3, No. 4, 1971, pp. 1338-
1351. doi:10.1103/PhysRevB.3.1338
[16] P. Scherrer, “Bestimmung der Grösse und Derinneren
Struktur von Kolloidteilchen Mittels Röntgenstrahlen,”
Nachrichten von der Gesellschaft der Wissenschaften zu
Göttingen, Mathematisch-Physikalische Klasse, Vol. 26,
No. 1, 1918, pp. 98-100.
[17] G. Yue, J. D. Lorentzien, J. Lin, D. Han and Q. Wang,
“Photoluminescence and Raman Studies in Thin-Film
Materials: Transition from Amorphous to Microcrystal-
line Silicon,” Journal of Applied Physical Letters, Vol. 75,
No. 4, 1999, pp. 492-494. doi:10.1063/1.124426
[18] D. Beeman, R. Tsu and M. F. Thorpe, “Structural Infor-
mation from the Raman Spectrum of Amorphous Sili-
con,” Physical Review B, Vol. 32, No. 2, 1985, pp. 874-
878. doi:10.1103/PhysRevB.32.874
[19] Y. He, C. Yin, G. Cheng, L. Wang, X. Liu and G. H. Hu,
“The Structure and Properties of Nanosize Crystalline Si-
licon Films,” Journal of Applied Physics, Vol. 75, No. 2,
1994, pp. 797-803. doi:10.1063/1.356432
[20] E. C. Freeman and W. Paul, “Optical Constants of rf
Sputtered Hydrogenated Amorphous Si,” Physical Review
B, Vol. 20, No. 1, 1979, pp. 716-728.