Structural properties of SnxSy thin films prepared by plasma-enhanced chemical vapor deposition

Article abstract of DOI:10.1149/1.1393962

The growth and structural properties of Sn_xS_y (SnS₂, Sn₂S₃) prepared by the plasma-enhanced chemical vapor deposition process have been studied systematically. Sn_xS_y thin films were prepared by the decomposition of H₂S and SnCl₄ vapors mixture in a capacitively coupled 13.56 MHz radio frequency glow discharge chamber with a radially symmetric flow pattern. Hydrogen was used as a diluent gas and for removing chlorine radicals generated by the SnCl₄ decomposition. The deposition pressure, substrate temperature, and plasma power density were kept constant at 50 mTorr, 150°C, and 25 mW/cm², respectively. The relative concentration of the precursor materials, g, defined as the ratio of tin chloride mass flow rate to the sum of tin chloride and hydrogen sulfide mass flow rates, was varied from 0 to 1.0. A total mass flow rate of 25 standard cubic centimeter per minute for the precursor materials and the diluent gas was used in all the cases. It was found that (i) For g < 0.2, the deposited thin films contain only the 2H-SnS₂ phase, and show a hexagonal crystalline structure with a preferential growth of the c axes perpendicular to the plane of the substrate; (ii) For 0.2 ≤ g < 0.5, and g > 0.6, the deposited films contain a mixture of hexagonal SnS₂ and orthorhombic Sn₂S₃ compounds; and (iii) For g close to 0.5, the deposited material has only the Sn₂S₃ compound. In all the samples, the size of the crystallites and the lattice parameters were estimated from the width of the X-ray diffraction peaks and the interplanar distances. It was found that the chemical composition, the crystallinity, and the preferential growth of the deposited material are well controlled by adjusting the value of g.

Full text of DOI:10.1149/1.1393962

3708
Journal of The Electrochemical Society, 147 (10) 3708-3717 (2000)
S0013-4651(00)02-060-7 CCC: $7.00 © The Electrochemical Society, Inc.
Structural Properties of SnxSy Thin Films Prepared by Plasma-Enhanced
Chemical Vapor Deposition
A. Sanchez-Juareza and A. Ortizb
aDepartment of Solar Materials, Energy Research Center, Universidad Nacional Autonoma de Mexico, Temixco, Morelos,
62580 Mexico
bMaterials Research Institute, Universidad Nacional Autonoma de Mexico, Coyoacan, D. F., 04510 Mexico
The growth and structural properties of SnxSy (SnS2, Sn2S3) prepared by the plasma-enhanced chemical vapor deposition process
have been studied systematically. SnxSy thin films were prepared by the decomposition of H2S and SnCl4 vapors mixture in a capac-
itively coupled 13.56 MHz radio frequency glow discharge chamber with a radially symmetric flow pattern. Hydrogen was used
as a diluent gas and for removing chlorine radicals generated by the SnCl4 decomposition. The deposition pressure, substrate tem-
perature, and plasma power density were kept constant at 50 mTorr, 150ЊC, and 25 mW/cm2, respectively. The relative concentra-
tion of the precursor materials, g, defined as the ratio of tin chloride mass flow rate to the sum of tin chloride and hydrogen sul-
fide mass flow rates, was varied from 0 to 1.0. A total mass flow rate of 25 standard cubic centimeter per minute for the precursor
materials and the diluent gas was used in all the cases. It was found that (i) For g < 0.2, the deposited thin films contain only the
2H-SnS2 phase, and show a hexagonal crystalline structure with a preferential growth of the c axes perpendicular to the plane of
the substrate; (ii) For 0.2 Յ g < 0.5, and g > 0.6, the deposited films contain a mixture of hexagonal SnS2 and orthorhombic Sn2S3
compounds; and (iii) For g close to 0.5, the deposited material has only the Sn2S3 compound. In all the samples, the size of the
crystallites and the lattice parameters were estimated from the width of the X-ray diffraction peaks and the interplanar distances.
It was found that the chemical composition, the crystallinity, and the preferential growth of the deposited material are well con-
trolled by adjusting the value of g.
© 2000 The Electrochemical Society. S0013-4651(00)02-060-7. All rights reserved.
Manuscript submitted February 15, 2000; revised manuscript received July 3, 2000.
process that meets these requirements. In order to use these materi-
als in a thin film device, it is important to know their structural prop-
erties and their dependence on the preparation conditions. To our
knowledge, only a few papers have been devoted to the study of the
structural properties of SnxSy polycrystalline thin film materials, and
in only one, the PECVD process was used. Recently Ortiz et al.18
have reported the deposition of p-type SnS polycrystalline thin films
by PECVD. The present work deals with the deposition of SnxSy thin
films by PECVD. Here, growth and structural properties are ana-
lyzed as a function of the concentration ratio of the source materials.
An optimization of the deposition parameters to grow SnS2 and
Sn2S3 thin film materials is also presented.
Binary compounds like SnS, SnS2, and Sn2S3, based on Sn and S,
are semiconductor materials1-3 with a high potential use in optoelec-
tronic devices due to their optical and electrical properties.4-6 These
semiconductors have been prepared by different processes such as
chemical melt growth,3,7-9 chemical vapor transport,10,11 physical
vapor transport,12 chemical vapor deposition (CVD),13 electroless
deposition,14 spray pyrolysis,15-17 and glow discharge.18 The struc-
tural, optical, and electrical properties of the different compounds in
the binary Sn-S system depend on the preparation technique.
Tin sulfide, SnS, crystallizes into a deformed NaCl structure,
which is described to be composed of tightly bonded double layers
of Sn and S atoms along the c axis, with van der Waals-type binding
between layers.19 The single crystalline material has shown indirect
Experimental
3
transitions with an optical bandgap of about 1.3 eV, which lies in
The chemical reaction.—From a mixture of SnCl4 and H2S
vapors diluted with H2, depending on the mole ratio of the reactants
and external stimulants, different sulfides of tin can be formed in the
following reactions
the optimum range for absorber materials to be used in solar cells
applications.20 It is a p-type material whose conductivity can be
increased by doping with Ag, Al, N, and Cl.3,21 On the other hand,
polycrystalline thin films have an optical bandgap in the range of 1.1
to 1.3 eV. 13
Tin disulfide, SnS2, is a layered compound that crystallizes in the
hexagonal CdI2-type structure with at least three basic polytypes that
are classified in terms of their stacking sequence.22 The basic struc-
ture unit represented by (A␥B)n is called 2H-SnS2, where ␥ corre-
sponds to Sn, and A and B correspond to the S layers. SnS2 single
crystals have shown an n-type electrical conductivity,23 and optical
bandgaps in the range of 2.12 to 2.44 eV. 7,10-11 SnIISnIVS3 is classi-
fied as a Type I mixed valence compound. This material shows semi-
conductor behavior with an energy bandgap dependent of its crys-
talline structure and stoichiometry. Thin film polycrystalline sam-
ples of Sn2Ϯx S3 present allowed indirect optical transitions with an
optical bandgap ranging from 1.16 to 1.9 eV. 13,15,16
k1
r
D SnS ϩ E HClD
[1a]
k2
r
A SnCl4 ϩ B H2S ϩ C H2
F SnS2 ϩ E HClD ϩ G H2 [1b]
k3
r
H Sn2S3 ϩ E HClD
[1c]
The values of A, B...H, vary with the reactants and products of the par-
ticular reaction and are related to the number of moles involved, while
k1, k2, k3 are the unknown reaction rate constants. The reactants in the
left side of Eq. 1 could produce the products in the right side if the
reaction is activated by means of external energy.
Thin film preparation.—SnxSy thin films were prepared in a glow
discharge system with a radially symmetric flow pattern. A 13.56
MHz radio frequency (rf) generator with a matching network drives
the glow discharge onto a capacitively coupled planar electrode con-
figuration. Figure 1 shows a schematic diagram of the system. To
control the gas flow rate of the precursor vapors (SnCl4 and H2S)
and the diluent gas (H2) electronic mass flow controllers were used.
The deposition pressure was measured by a baratron and it was con-
trolled automatically with a throttle valve. Two kinds of substrates
Taking into account the optoelectronic properties of Sn and S-
based compounds, it can be inferred that these materials could be
used to build photovoltaic p-n or p-i-n structures with a conversion
efficiency of about 25%.20 These structures should be low cost
devices because the materials involved are inexpensive, nonstrategic,
and abundant in nature. However, in order to create a cost-competi-
tive photovoltaic cell, it is necessary to have a precisely, controllable,
simple, and inexpensive large area thin film deposition technique.
The plasma-enhanced chemical vapor deposition (PECVD), is a
Downloaded on 2014-06-01 to IP 155.33.16.124 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 147 (10) 3708-3717 (2000)
3709
S0013-4651(00)02-060-7 CCC: $7.00 © The Electrochemical Society, Inc.
plasma, as a function of the gas pressure in the chamber is shown in
Fig. 2. The grounded plate was at room temperature in all cases. For
each gas there exists a minimum value for the rf power that depends
on the gas pressure. This value is related to the minimum energy
required to the breakdown. This behavior can be explained consid-
ering that at low pressure the electron mean free path (␭e) is large,
and thus the majority of the electrons reach the electrodes without
colliding with gas molecules. Therefore, the lower the pressure, the
higher is the value of rf power needed to produce the breakdown of
the gas. On the other hand, ␭e becomes short at higher pressures. In
this case, electrons can not gain a high enough energy from the ac
electric field to ionize the gas because of their frequent collisions
with the gas molecules. Thus, rf power increases as the pressure
increases. By analogy with the Paschen curve of a dc plasma,25 the
left side of the minimum in Fig. 2 is called the low pressure regime
and the right side, the high pressure regime. The rf power needed to
sustain the plasma is lower than the one needed to the breakdown of
the plasma. This is explained by the fact that once the plasma starts,
electrons and ions are trapped in the discharge space. These elec-
trons frequently cause ionization of gas molecules by inelastic colli-
sions, then the rf power needed to sustain the plasma decreases from
its initial value to a value that keeps the rate of generation and
recombination of electrons balanced.25 Figure 2 also shows that H2S
requires less rf power for starting the glow than the others, in range
of pressure considered. This behavior could be understood consider-
ing that the energy needed to produce an excitation state in a mole-
cule is lower than the one required to produce an ionization state.
Since the ionization potential of the H2S molecule (10.45 eV) is
lower than those of SnCl4 (11.88 eV) and H2 (15.42 eV) mole-
cules,26 it can be expected that the molecule that requires less ener-
gy for having an ionization state will require less energy for having
an excited state. Thus, the H2S molecule has an excitation state at
relatively less rf power than the others.
Figure 1. Schematic diagram of the PECVD chamber with a radially sym-
metric flow pattern. A fixed distance of 3 cm separates the two circular elec-
trode plates with a diam of 16 cm.
were used: Pyrex glass slices, chemically and ultrasonically cleaned
and 200 ⍀ cm n-type (100) single crystal silicon wafers, both with
1.5 ϫ 2.5 cm area. The flow rates of the precursors were chosen to
meet a maximum flow rate of 5 standard cubic centimeter per minute
(sccm) for the SnCl4 and H2S mixture. The liquid SnCl4 vapor
source was kept at 40ЊC in all cases. The substrate temperature was
controlled by means of a temperature controller at a fixed value of
150ЊC through a K-type thermocouple. The plasma power density
and the process pressure were kept constant at 0.025 W/cm2 and
50 mTorr, respectively. The concentration rate of the precursors, g,
defined by the following relation
In order to know if there are reactions besides excitation and de-
excitation processes, the solid materials deposited onto glass sub-
strates in the low pressure regime of Fig. 2 for SnCl4 alone, the
SnCl4/H2 mixture and H2S discharges were investigated under the
following conditions: process pressure of 50 mTorr; rf power of 5 W
(power density of 25 mW/cm2), and deposition time of 30 min. Two
substrate temperatures were used: 35 and 150ЊC, to analyze the pos-
sible effects of this parameter on the decomposition of the precursors.
[SnCl4]
g ϭ
[2]
[SnCl4] ϩ [H2S]
was varied from 0.0 to 1.0 in 0.1 steps. The square brackets in Eq. 2
indicate the flow rate in sccm. H2 was added at a flow rate of
20 sccm as a diluent gas. The total gas flow rate for H2, SnCl4, and
H2S was 25 sccm in all cases.
The SnCl4 discharge.—Figure 3a shows the XRD patterns of the
solid materials deposited at 35 and 150ЊC, showing diffraction peaks
Film characterization.—The thickness of the deposited films on
glass was measured with an Alpha Step 100 profilometer. The crys-
tal structure was analyzed by X-ray diffraction (XRD) measure-
ments on a Siemens D-500 diffractometer using Cu K␣ radiation of
␭ ϭ 1.5405 Å. The diffraction patterns were taken with a scan rate
of 2␪ ϭ 0.5/min. Surface morphology was observed by scanning
electron microscopy (SEM) using Cambridge-Leica Stereoscan 440
SEM equipment. The composition of the film was determined by the
energy dispersive spectroscopy (EDS) analysis using an Oxford sys-
tem detector attached to the scanning electron microscope.
Results and Discussion
Primary processes.—In a glow discharge process, the electrons
are the principal sources for transferring electrical energy to the gas
molecules through elastic and inelastic collisions that lead and con-
tribute to a variety of chemical reactions.24 Among them, the excita-
tion and deexcitation processes produce the glow. Hence, it is impor-
tant to know the gas pressure and electric power to be used in order to
produce the desired gas discharge breakdown in the reaction chamber.
The physical conditions necessary to produce and sustain a glow
discharge plasma using either SnCl4, H2S, or H2 for the chamber
shown in Fig. 1 were investigated separately. The flow rates of SnCl4
and H2S vapors were fixed at 5 sccm, while that of H2 was fixed at
20 sccm. The minimum rf power required for the breakdown of the
Figure 2. Behavior of the gas discharge breakdown for the vapor precursors
and H2. Tin tetrachloride and hydrogen sulfide are fed down into the cham-
ber at a rate of 5 sccm, while hydrogen is fed at 20 sccm.
Downloaded on 2014-06-01 to IP 155.33.16.124 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

3710
Journal of The Electrochemical Society, 147 (10) 3708-3717 (2000)
S0013-4651(00)02-060-7 CCC: $7.00 © The Electrochemical Society, Inc.
SnCl4(g) ϩ 2H2 } Sn(s) ϩ 4HCl(g)
[5]
*
Also, the H2 species produced by the plasma like H2 (excited mole-
cules), H (atoms), H* (excited atoms), Hϩ2 (ions), and Hϩ (protons),24
could be responsible for this overall chemical reaction in chain reac-
tions of the type expressed in Eq. 6
H ϩ Cl2 r HCl ϩ Cl
Cl ϩ H2 r HCl ϩ H
HCl r H ϩ Cl
[6]
Furthermore, for producing Sn Sy thin films without chlorine
incorporation, hydrogen added toxthe reaction could remove Cl
atoms from the gas phase and even from the growing film. It is seen
from Eq. 1 that the presence of H2 in the mixture reduces Sn4ϩ to
Sn2ϩ. In addition, it can combine with the Cl present in the system
to form HCl. Thus, the extraction of Cl during the deposition process
of these thin films can be controlled by the amount of hydrogen
added to the reaction chamber. It is important to note that an excess
of H2 per each SnCl4 molecule is necessary for extracting as much
chlorine as possible. This behavior has been studied for chlorinated
SiO2 thin films prepared by PECVD. In this case Alonso et al.27
found that it is necessary to add a H2 flow rate more than twice that
of the precursor flow rate, in order to eliminate Cl incorporation in
the film. In the present work, it is expected that Cl incorporation will
be avoided since the flow rate of H2 to that of the precursor was set
at four.
Figure 3. XRD patterns of a material deposited on a glass substrate using a
rf power of 5 W, substrate temperatures of 35 and 150ЊC, and plasma gener-
ated by vapors of SnCl4. The upper pattern corresponds to a material deposit-
ed when hydrogen was used as a diluent in a proportion 1:4 of SnCl4/H2.
attributable to SnCl4, SnCl4и5H2O, and SnCl2и2H2O (JCPDS no. 31-
1394, 31-1395, and 32-1380, respectively), all polycrystalline mate-
rials. No other compound could be identified with these peaks, and
no peak due to elemental Sn was identified in the patterns. Thus, at
these conditions, SnCl4 is not decomposed into Sn and Cl radicals.
The chemical processes that could take place due to electron-mole-
cule collisions, but not limited to them, are represented by (r) in
Eq. 3, where e denotes the electron. These processes suggest that a
significant amount of free Cl, Cl2, and different tin chloride radicals
will be produced during the deposition as by-products of the disso-
ciation of SnCl4
When a high power density, i.e., 500 mW/cm2, was used to pro-
duce the glow discharge, it was found that the deposited material has
incorporated compounds related to SnCl2 and metallic tin, hence the
SnCl4 is decomposed into Sn and Cl radicals.
The H2S discharge.—For the substrate at a temperature of 35ЊC, the
deposited thin film showed a yellow color in reflection that could be
associated with solid sulfur formed by the decomposition of H2S.28
There was no film deposition at a substrate temperature of 150ЊC.
This can be explained as follows: at room temperature, the generat-
ed sulfur is condensed and adsorbed on the substrate surface with a
residence time long enough to form a thin film. At 150ЊC, the ad-
sorbed sulfur acquires the necessary energy to be desorbed with a
high desorption rate. Thus, sulfur is returned to the plasma generat-
ing more active species. In the case of the material deposited at low
temperature, the XRD pattern did not show any diffraction peak that
could be attributed to sulfur. This may be due to the small thickness
of the film or perhaps due to the amorphous nature of the deposited
material. The overall reaction28 that could take place is given by
Dissociation:
SnCl4 ϩ e r SnCl2 ϩ Cl2 ϩ e
Cl2 ϩ e r 2Cl ϩ e
*
Excitation:
SnCl4 ϩ e r SnCl4 ϩ e
*
Cl2 ϩ e r Cl2 ϩ e
[3]
Cl ϩ e r Cl* ϩ e
SnCl4 ϩ e r SnClϩ4 ϩ 2e
Cl2 ϩ e r Clϩ2 ϩ 2e
Cl ϩ e r Clϩ ϩ 2e
Ionization:
H2S } H2 ϩ (1/8) S8
[7]
caused by one or more of the following processes, represented by
(r) in Eq. 8, derived from electron-molecule collisions, but not lim-
ited to them
Further experiments like mass spectrometry and optical emission
spectroscopy should be carried out in order to know the ionic and
luminescent species that can be associated with the possible reactions.
Another reaction that can be associated with the formation of a solid
film is the condensation of SnCl4 vapors onto the glass substrate via
the reaction13
Excitation:
H2S ϩ e r H2S* ϩ e
*
H2 ϩ e r H2 ϩ e
H ϩ e r H* ϩ e
S ϩ e r S* ϩ e
SnCl4(g) } SnCl2(s) ϩ Cl2(g)
[4]
Dissociation:
Ionization:
H2S ϩ e r H ϩ SH ϩ e
H ϩ SH ϩ e r 2H ϩ S ϩ e
On the other hand, this experiment did not show evidence that the
substrate temperature enhances the SnCl4 decomposition in the tem-
perature range considered.
[8]
Sm ϩ e r mS ϩ e
The SnCl4/H2 discharge.—When hydrogen was used as a diluent, in
a proportion 1:4 of SnCl4/H2, the material deposited was only metal-
lic tin as indicated by the XRD pattern in Fig. 3b. This result can be
explained on the basis of hydrogen being considered as a strong
reducing agent. Then, the following overall reaction takes place
H2S ϩ e r H2Sϩ ϩ 2e
H2 ϩ e r Hϩ2 ϩ 2e
S ϩ e r Sϩ ϩ 2e
Downloaded on 2014-06-01 to IP 155.33.16.124 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 147 (10) 3708-3717 (2000)
3711
S0013-4651(00)02-060-7 CCC: $7.00 © The Electrochemical Society, Inc.
Thus, in the low pressure regime, H2S is decomposed due to
inelastic collisions with electrons in the plasma. However, SnCl4 is
not decomposed by the electron-molecule collision.24 This result is
a direct consequence of the energy required to break a chemical
bond. The H–S bond needs 3.57 eV, while Cl-Sn needs 4.29 eV. 26
Deposition of SnxSy thin films.—Figure 4 shows the behavior of
the rf power against pressure to establish the breakdown plasma
process in a 1:1 mixture of SnCl4 and H2S, with and without H2 dilu-
tion. In both cases, there is a minimum in the curve that defines the
concept of low pressure and high pressure regime, in agreement with
Paschen’s curve. It can be seen that H2 enhances the breakdown at
pressures higher than 150 mTorr, in a similar way as is shown in the
inset of Fig. 2; while at pressure lower than 150 mTorr, the mixture
of H2S and SnCl4 breaks down more readily.
In the present study, all of the deposited SnxSy thin films were
carried out using deposition parameters in the low pressure regime.
Therefore, there is no primary decomposition of SnCl4 by the plas-
ma and the chemical reactions that lead to thin film growth must be
related to the H2S decomposition and the reduction of SnCl4 by H2.
Deposition rate.—Figure 5 shows the behavior of the deposition
rate (Rd), defined by the thickness divide by deposition time, as a
function of g, for samples prepared at a deposition time of 30 min.
Each point represents the mean value of Rd for at least six samples
prepared at the same conditions and located at the same position on
the grounded electrode. The error bars in each point show the exper-
imental standard deviation for the Rd mean value. In this figure, the
g range is divided into five regions. In regions I and II, Rd increases
as g increases. In region III, Rd decreases up to a minimum value for
g Ϸ 0.6; after which, Rd increases as g increases up to about g ϭ 0.7.
Region IV shows a saturation behavior of Rd with g; and finally in
region V when g ϭ 1, Rd drops to a very low value. To explain this
behavior it is necessary to analyze the deposited material. Thus,
XRD and EDS analyses were performed in samples with similar
thickness (Х0.15 ␮m ).
Figure 5. Behavior of the deposition rate as a function of g. Regions I, II, III,
IV, and V are defined by the intervals shown on the top of this figure.
common feature for single crystal SnS2 prepared by chemical vapor
transport.30 To our knowledge, no one else has reported this kind of
growth for polycrystalline SnS2 thin films. Prominent peaks, the cal-
culated grain size, and their corresponding d values for these SnS2
thin films are listed and compared with those of the standard patterns
in Table I. The 2␪ peak position in each pattern is shifted to lower
values as g increases. This behavior could be attributed to the inclu-
sion of chlorine impurities in the SnS2 lattice or to slight departures
from the stoichiometry of the compound, i.e., a metal to sulfur ratio
greater than 1:2. Both hypotheses are supported later. The lattice
parameters found are in close agreement with the reported values for
the standard powder samples of SnS2, and they are similar to those
reported for single crystals of SnS2 prepared by other processes (see
for instance, Kourtakis et al.11 and Ray et al.31). It can be seen from
Table I that the lattice parameter c experiences a slight increase as g
increases, and its mean value is 1.2% higher than the standard one.
This small discrepancy is also associated with the same cause that
has been proposed to explain the small shift in the 2␪ peak position
observed in the patterns of Fig. 6. Therefore, in region I, an increase
in g does not affect the preferential growth of the films and the lat-
tice parameters of the SnS2 crystal. On the other hand, the magni-
Structural and compound identification.—Figure 6 shows the
XRD patterns of typical samples prepared for g < 0.2 (region I).
Peak identification was performed following the procedure de-
scribed in Ref. 29. The observed peaks match with those of the stan-
dard pattern for tin disulfide (2H-SnS2: JCPDS no. 23-677). This
compound has an hexagonal structure lying in the space group
P3m1. The strong peak located at 2␪ around 15Њ is due to reflections
from the (001) plane. The appearance of the [00l] strong peak indi-
cates that the deposited material is highly oriented with the c axes
perpendicular to the plane of the substrate. This type of growth is a
Figure 4. Gas discharge breakdown behavior for a mixture 1:1 of SnCl4 and
H2S: curve (ᮀ) without hydrogen dilution, and curve (᭹) with hydrogen
dilution.
Figure 6. Typical XRD patterns for samples prepared using g Յ 0.2. All the
patterns show a strong peak for the (001) planes, indicating preferred orien-
tation along [001].
Downloaded on 2014-06-01 to IP 155.33.16.124 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

3712
Journal of The Electrochemical Society, 147 (10) 3708-3717 (2000)
S0013-4651(00)02-060-7 CCC: $7.00 © The Electrochemical Society, Inc.
Table I. Prominent peak positions (2␪ values) of XRD peaks, corresponding d-values, gs, lattice parameters, and their identification as 2H-
SnS2 for samples prepared with g < 0.2.
Comparison
Observation
SnxSy polycrystalline thin films
2H-SnS2 (JCPDS data file 23-677)
Systematic hexagonal; Lattice parameters:
a ϭ 3.649 Å; c ϭ 5.899 Å
(Strong peak positions)
Lattice parameters
g
Phase
Peak
d-value
I/Io
gs
Peak
d-value
I/Io
(hkl)
(2␪)
(Å)
(nm)
a
b
c
(2␪)
(Å)
15.07
30.22
5.879
2.960
100
112
15.029
30.262
5.891
2.951
100
115
(001)
(002)
0.09
0.1
SnS2
SnS2
SnS2
SnS2
16
14
13
13
5.913
5.966
5.981
6.002
14.83
29.85
5.969
2.991
100
112
15.029
30.262
5.891
2.951
100
115
(001)
(002)
14.81
28.31
5.981
3.148
100
113
15.029
28.199
5.891
3.162
100
1130
(001)
(100)
0.12
0.19
3.639
3.637
14.76
28.35
6.002
3.145
100
117
15.029
28.199
5.891
3.162
100
1130
(001)
(100)
prepared using a deposition pressure of 50 mTorr, which is below the
threshold value found in Fig. 4, and there is no evidence of the SnS
compound in all the deposited films. This difference could be attrib-
uted to the deposition pressure that gives another reaction path. The
deposition pressure is an important parameter that must be taken into
consideration. Similar results have been observed for a-Si:H and a-
SiC:H thin films produced by PECVD.32,33
As g increases, taking values in region III, Sn2S3 apparently stops
growing and the SnS2 compound grows again with the hexagonal
crystal structure, in a similar way as those prepared using g values
within the interval (0.12, 0.31). This is evident from Fig. 7b in which
the XRD patterns corresponding for samples grown using g ϭ 0.56
and 0.6 show strong peaks corresponding to the SnS2 compound,
meanwhile the peaks corresponding to the Sn2S3 almost disappear.
There are other weak peaks in the patterns shown in Fig. 7, but the
signal-to-noise ratio of those peaks is very small. All the samples
prepared within the interval (0.2, 0.49) do not show any preferential
growth like that observed for the SnS2 compound. Prominent peaks,
corresponding d values, grain size (gs), lattice parameters, and their
respective assignments for samples prepared within 0.2 < g Յ 0.6
are listed and compared in Table II.
tudes of the reflection peak at 2␪ ϭ 15.03Њ decreases as g increases
(see Fig. 6). The crystallite size calculated using these peaks
decreases slightly as g increases, showing a trend toward a mean
value of 13 nm (see Table I). This reduction could also be associat-
ed to the inclusion of chlorine impurities at the crystal network in the
sulfur position.
Figure 7 shows the X-ray diffractograms for typical samples pre-
pared at g values corresponding to region II (Fig. 7a) and to region III
(Fig. 7b). Peak positions of each one of these patterns were compared
with the standard JCPDS data file card no. 23-677 for 2H-SnS2 and
card no. 14-619 for Sn2S3. Figure 7a shows peaks that correspond to
the 2H-SnS2 and Sn2S3 compounds. Therefore, the identification pro-
cedure confirmed that both compounds grew simultaneously in this
interval of g, while only the Sn2S3 is deposited when g takes a value
close to 0.49. This fact is confirmed in Fig. 7b in which all the peak
positions in the spectra corresponding to g ϭ 0.49, except the two at
2␪ Х 24.9 and 40.3Њ, match the standard for the Sn2S3 compound.
The peaks at 2␪ Х 24.9 and 40.3Њ are associated with the SnCl2и2H2O
compound (see Fig. 3a). Using a similar chamber and under similar
deposition conditions, except for the deposition pressure (g ϭ 0.5 and
a deposition pressure of 150 mTorr), Ortiz et al.18 found that the
deposited material was SnS. In the present work, all the samples were
Figure 8 shows the XRD patterns for samples prepared using g
values in region IV. There are big peaks corresponding to 2H-SnS2
and SnCl2и2H2O, and small peaks located around 2␪ Х 12.8Њ and
38Њ that corresponded to the Sn2S3 (see Fig. 8a). In Fig. 8b, there are
peaks other than those which could be associated with the 4H-SnS2
phase (polytype of the 2H-SnS2 with double c axes), ␤-S, and metal-
lic tin (JCPDS no. 21-1231, 34-0941, and 4-673, respectively). The
last assignment was done because for g ϭ 1, the deposited material
was metallic tin (see Fig. 3b). The peak corresponding to 4H-SnS2
at 2␪ Х 29.2Њ was only observed for g > 0.72, means that as g
increases, the 2H is gradually transformed into the 4H structure. The
reverse behavior has been observed for tin disulfide single crystal
grown by chemical vapor transport.30 Prominent peaks and their cor-
responding d values for the deposited thin films prepared using g
values within the considered interval (0.6 < g Յ 1) are listed and
compared in Table III. There is a small shift in all the diffraction
peak positions with respect to the standard peak position considered
in Tables II and III. This behavior is due to the fact that both SnS2
and Sn2S3 compounds are grown along with SnCl2. Tables II and III
also show that the crystalline gs for the 2H-SnS2 phase, calculated
using the peak at 2␪ Х 15Њ, almost remains constant with a mean
value of 12.8 nm for all g values except for g > 0.75 for which it is
of the order of 28 nm. The latter is calculated using the peak at 2␪ Х
29.2Њ. This difference in the crystalline grain size can be associated
with the change in the size of the lattice parameters, because the
deposited material is identified as the 4H-SnS2 phase. The Sn2S3
Figure 7. XRD patterns for samples prepared using discrete g values from
0.22 up to 0.6. These patterns show how the crystal structures are changing
as g increases.
Downloaded on 2014-06-01 to IP 155.33.16.124 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 147 (10) 3708-3717 (2000)
3713
S0013-4651(00)02-060-7 CCC: $7.00 © The Electrochemical Society, Inc.
Table II. Prominent peak positions (2␪ values) of XRD peaks, corresponding d-values, gs, lattice parameters, and their respective
identification, for samples prepared with 0.2 < g Յ 0.6.
Comparison
2H-SnS2
(JCPDS data file 23-677)
Systemic Hexagonal;
Lattice parameters:
Comparison
Sn2S3 (JCPDS data file 14-619)
Systemic Orthorhombic;
Lattice parameters:
a ϭ 8.864 Å; b ϭ 14.020 Å;
Observation
SnxSy polycrystalline thin films
(Strong peak positions)
a ϭ 3.649 Å; c ϭ 5.899 Å
c ϭ 3.747 Å
Lattice parameters
g
Phase Peak d-value I/Io
(2␪) (Å)
gs
(nm)
Peak d-value I/Io (hkl)
Peak d-value I/Io (hkl)
a
b
c
(2␪)
(Å)
(2␪)
(Å)
SnS2
14.89 5.932 100 12
Sn2S3 31.14 2.875 100 14
SnS2 14.98 5.921 100 13
Sn2S3 21.66 4.094 100 17
SnS2 28.36 3.144 100 13
Sn2S3 31.90 2.804 100 20
SnS2 28.37 3.143 100 13
Sn2S3 31.82 2.811 100 19
3.638
—
5.932 15.029 5.891 100 (001)
—
30.916
—
2.89
—
—
0.22
0.31
0.40
0.44
8.653 13.836 3.847
3.638 5.921 15.029 5.891 100 (001)
8.778 13.885 3.757
3.637 5.806 28.199 3.162 130 (100)
8.923 13.754 3.738
3.629 5.80 28.199 3.162 130 (100)
—
—
—
-
115 (310)
—
—
21.498
—
4.13
—
—
—
—
—
—
100 (130)
—
—
31.936
—
2.80
—
—
—
—
—
—
115 (211)
—
8.888 14.142 3.73
—
—
—
—
31.936
31.936
2.80
2.80
115 (211)
115 (211)
0.49 Sn2S3 32.11 2.785 100 28
0.60 SnS2 14.73 6.012 100 13
8.853 13.907 3.709
3.638
6.012 15.029 5.89 100 (001)
compound did not show any preferential growth and the crystalline
grain size calculated at the peak position considered shows an in-
creasing trend as g increases up to 0.69, after that, the grain size
tends to a value of 30 nm.
shown in Fig. 9 have been found also for SnxS thin film prepared by
low temperature chemical precipitation, dip deposition, and cathod-
ic electrodepositon.13,31,34
The composition of the deposited SnxSy thin films, determined by
EDS analyses, can be used to determine the stoichiometry of the
compound. The atomic percentage (atom %) of Sn, S, and Cl incor-
porated in the deposited material as a function of g are shown in
Fig. 10. It can be seen that for g < 0.2 and for g ϭ 0.6, the atomic
percentage of Sn and S incorporated are almost in the ratio of Sn/S ϭ
0.5, thus the stoichiometry of the thin films was found to be 1:2,
which confirms the formation of the SnS2 compound. This is in
agreement with the compound identified by XRD. In the case of g ϭ
0.09, it was found that the ratio Sn/S Х 0.17, meaning that the
deposited material has a significant Sn deficiency. Thus, the stoi-
chiometry is of the form Sn1ϪxS2. On the other hand, it was not pos-
sible to define a unique stoichiometry for 0.2 < g < 0.95, except for
0.6, because the deposited material is formed by two compounds.
The quantity of Sn incorporated remains almost constant with a value
around 40 atom %, while the quantity of S is almost 55 atom %, then
for g ϭ 0.49 the ratio 40 atom % Sn to 55 atom % S is approximate-
ly in the ratio 2:3, that identifies the Sn2S3 compound. Also, it can be
seen that there is chlorine incorporation in the deposited material for
all the g values, except for g < 0.1.
Surface morphology and composition.—Figure 9 shows typical
SEM micrographs of SnxSy thin films prepared using different g val-
ues. The surface appearance changes from smooth to rough as g
increases. For g > 0.2, the micrographs show crystal grains com-
prised of randomly oriented wormlike structures or fibrils. The fib-
rils appeared to be larger for g > 0.6 than those for g < 0.6. This sur-
face appearance could be related to the crystalline phase of the mate-
rial: SnS2 for g < 0.2 and a mixture of SnS2 and Sn2S3 compounds
with inclusion of crystallites of SnCl2 for g > 0.2. A very smooth sur-
face is obtained for the SnS2 compound, while a rough surface is
observed for the mixture. The latter could be attributed to the growth
competition between the SnS2 and Sn2S3 compounds, and the depo-
sition of SnCl2 occurring for g > 0.2. Similar features, like those
The atom % of Cl shows a small increasing trend as g increases,
and its maximum concentration is less than 5 atom % even for g ϭ 1.
There are two ways in which Cl can be incorporated in the deposited
material: as a tin chloride related radical, as a SnCl2-like compound
as has been identified in XRD (see Fig. 7 and 8), or as an impurity
replacing sulfur. The inclusion of atomic Cl in the deposited material
as an impurity instead of sulfur (the ionic radius of chloride is less
than that of the sulfide), or as a compound, could be responsible for
the shift of the peak position of the XRD patterns shown in Fig. 6, 7,
and 8, as well as for the deviation of the values in the lattice parame-
ters of the deposited material.29 In the case of SnS2, the Cl incorpo-
ration could affect the electrical properties of the films, as has been
shown in previous works.3,21 Although the Cl concentration found is
small, the quantity of H2 added to the reaction or the deposition con-
ditions was not enough to leave a chlorine free SnxSy network.
Formation process.—It was found that the precursor concentra-
tion, represented by g, affects the chemical composition of the
deposited material, giving SnS2 thin films for g < 0.2, and thin films
with a mixture of Sn2S3 and SnS2 for g > 0.2. In order to explain how
Figure 8. XRD patterns for samples prepared with g Ն 0.6. The peak at 2␪ Х
29.2Њ confirms the gradual transformation of the 2H-SnS2 into the 4H-SnS2
structure.
Downloaded on 2014-06-01 to IP 155.33.16.124 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

3714
Journal of The Electrochemical Society, 147 (10) 3708-3717 (2000)
S0013-4651(00)02-060-7 CCC: $7.00 © The Electrochemical Society, Inc.
Table III. Prominent peak positions (2␪ values) of XRD peaks, corresponding d-values, gs, lattice parameters, and their respective
identification for samples prepared with g > 0.6.
Comparison
Comparison
2H-SnS2
(JCPDS data file 23-677)
Systematic hexagonal;
Lattice parameters:
a ϭ 3.649 Å; c ϭ 5.899 Å
Sn2S3
(JCPDS data file 14-619)
Systematic Orthorhombic;
Lattice parameters:
Observation
SnxSy polycrystalline thin films
a ϭ 8.864 Å; b ϭ 14.020 Å;
(Strong peak positions)
c ϭ 3.747 Å
Lattice parameters
)
g
Phase
SnS2
Peak
(2␪)
d-value I/Io
(Å)
gs
(nm)
Peak d-value I/Io
(2␪) (Å)
(hkl
Peak d-value I/Io
(2␪) (Å)
(hkl)
a
b
c
15.04 5.894 100 12
3.645
—
5.894 15.029 5.89 100 (001)
0.69
0.72
0.75
Sn2S3 26.79 3.324 100 46
SnS2 15.10 5.866 100 13
Sn2S3 21.77 4.086 100 25
SnS2 14.99 5.910 100 13
Sn2S3 21.73 4.088 100 31
8.849 13.895 3.711
3.604 5.866 15.029 5.89 100 (001)
8.754 13.903 3.733
3.646 5.910 15.029 5.89 100 (001)
8.763 13.841 3.774
—
—
—
—
23.586 3.35
15 (111)
—
—
—
—
—
21.498 4.13 100 (130)
21.498 4.13 100 (130)
—
Sn2S3
4H-SnS2
(JCPDS data file 14-619)
Systematic Orthorhombic;
Lattice parameters:
(JCPDS data file 21-231)
Systematic Hexagonal;
Lattice parameters:
a ϭ 8.864 Å; b ϭ 14.020 Å;
a ϭ 3.645 Å; c ϭ 11.802 Å
c ϭ 3.747 Å
Lattice parameters
g
Phase
SnS2
Peak
(2␪)
d-value I/Io
(Å)
gs
(nm)
Peak d-value I/Io
(2␪) (Å)
(hkl)
Peak d-value I/Io
(2␪) (Å)
(hkl)
a
b
c
29.19 3.059 100 28
3.623
—
11.790 29.257 3.05
16 (101)
16 (101)
0.82
0.95
Sn2S3 21.86 4.065 100 27
SnS2 29.22 3.062 100 29
Sn2S3 21.82 4.074 100 28
8.203 14.04 03.734
21.498 4.13 100 (130)
21.498 4.13 100 (130)
3.614
11.790 29.257 3.05
8.943 13.909 13.689
Sn (JCPDS data file 4-673)
Systematic Tetragonal;
Lattice parameters:
a ϭ 5.831 Å; c ϭ 3.182 Å
Lattice parameters
g
Phase Peak d-value I/Io
gs
Peak d-value I/Io (hkl)
(2␪)
(Å)
(nm)
a
b
c
(2␪)
(Å)
1.0 Sn
32.0
2.795 100 18
5.831
3.225 32.018 2.793 90 (101)
the formation of these compounds occurs, it is supposed that all the
molecules that are fed down to the chamber undergo a chemical
reaction. Since all the processes were carried out in the low pressure
regime, the primary decomposition products are radicals related with
H2S, probably atomic and molecular sulfur, meanwhile, the SnCl4
vapor reacts with H2 and with those radicals generated by the H2S
decomposition mainly via ionic reactions. During the glow dis-
charge of the gases SnCl4, H2S, and H2, some of the processes pre-
Figure 9. SEM micrographs showing the
morphology of the SnxSy thin film com-
pounds deposited at different g values. It is
clear that the surface morphology changes
from smooth to rough as g increases.
Downloaded on 2014-06-01 to IP 155.33.16.124 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 147 (10) 3708-3717 (2000)
3715
S0013-4651(00)02-060-7 CCC: $7.00 © The Electrochemical Society, Inc.
tain number of them react with SnCl4 to produce the above men-
tioned reactions. Since XRD patterns for g < 0.2 did not show evi-
dence of atomic sulfur in the deposited thin film, there must exist
another reaction path involved during the process besides the one
represented in this equation. Thus, it was assumed that H2S is gen-
erated during the glow discharge, after the initial dissociation of the
inlet H2S through the following reactions
S r (1/8) S8
H2 ϩ (1/8) S8 r H2 ϩ Sm
H2 Sm r H2S ϩ SmϪ1
[10]
This assumption is supported by Traus et al.,36 These authors
showed the production of H2S from the dissociation of a mixture of
H2/H2S in a high pressure glow discharge. This process occurs when
the relative concentration of H2 in the mixture is higher than that of
the H2S by a factor of 2.33. In our case, the H2/H2S ratio is higher
than this value. Therefore, although the concentration of H2S is
higher than the concentration of SnCl4, it is possible to suppose that
only 50% of the molecules of the hydrogen sulfide are involved in
the reaction to produce the SnS2 material, while the rest react in the
way shown in Eq. 10. It was also found that for g Х 0.6 the deposit-
ed material shows mainly the SnS2 compound. Under this condition,
before starting the glow discharge, the atomic concentration of Sn is
1.5 times that of S. Then it is expected that the following reaction
takes place
Figure 10. Sn, S, and Cl atomic concentrations incorporated at the thin films
as a function of the precursor concentration.
sented in Eq. 3 to 8 will produce the products and reactions repre-
sented in Table IV. Thus, the SnxSy compound formation could be
explained on the basis of the processes described below.
SnS2 formation.—When g Յ 0.2, the atomic concentration of S
inside the reaction chamber before starting the glow discharge is
higher than that of Sn. For example, for g ϭ 0.2, for each atom of tin
there are four atoms of S, and the concentration of S increases as g
decreases. This argument holds because it is well known that the
mass flow rate measured in sccm is proportional to the number of
molecules per minute that is flowing to the reaction chamber. The
relationship is given by 1 sccm ϭ 2.69 ϫ 1019 molecules per
minute.35 Following the work by Engelken et al.,13 for growing thin
films of tin disulfide, it is necessary to have one Sn4ϩ ion and 2S2Ϫ
ions to produce Reaction d in Table IV. Because Sn2ϩ could be
formed by the reduction of SnCl2 with H2, the Sn4ϩ ion is obtained
by Reaction b in Table IV. Then, if all the SnCl4 molecules produce
the Sn4ϩ ions, there are more than twice as many atoms of S in the
reaction chamber, for producing the S2Ϫ ion to form the SnS2 com-
pound. The overall chemical reactions yielding the formation of
SnS2 thin films could be the following
3SnCl4 ϩ 2H2S ϩ 2H2 r SnS2 ϩ 2SnCl2 ϩ 8HCl
[11]
Though according to the stoichiometry of the reaction, only 10% of
H2 introduced to the chamber is required for the reaction shown in
Eq. 11, it was observed that thin films of SnS2 are formed only when
an excess of H2 is present. Equation 11 explains the presence of
SnCl2 in the deposited thin film. The XRD pattern for the samples
prepared with g ϭ 0.6 (see Fig. 7b) shows peaks attributed to
SnCl2и2H2O. Water molecules are attached to SnCl2 because this
compound absorbs water when the deposited thin films are exposed
to the atmosphere.
Sn2S3 formation.—In the presence of S2Ϫ and SnCl2 molecules, the
Sn2S3 compound can be formed through Reaction c in Table IV,
requiring four molecules of SnCl2 with three molecules of S to pro-
duce one molecule of Sn2S3 and two molecules of SnCl4.13 Regard-
ing the effect of the concentration of the precursors on the structure
and composition of the films (see Table II), the Sn2S3 compound was
obtained only in the case of samples prepared with a g value around
0.5, where the atomic concentrations of S and Sn in the reaction
chamber are the same before starting the glow. If it is assumed that
all the SnCl4 molecules are transformed to SnCl2, then only 75% of
the atomic sulfur generated by the decomposition of the H2S mole-
cules is necessary for the chemical balance of Reaction c on
Table IV. Under this condition, out of the 100% of the SnCl2 mole-
cules that are involved in this chemical reaction, 50% react with S to
1 SnCl4 ϩ 2 H2S r SnS2 ϩ 4 HCl
2 [H2S r H ϩ SH]
H ϩ SH r S ϩ 2H
2H r H2
[9]
The H2 molecules flowing inside the reaction chamber at a flow rate
of 20 sccm are not represented in these balanced equations, but a cer-
Table IV. Precursors, by-products, and reactions that could take place in a PECVD process using SnCl4 and H2S compound like precursors
for growing SnxSy thin film materials.
Precursors and diluent
SnCl4
By-products
Reactions in the presence of hydrogen and/or sulfur
SnCl2, Cl2;
Sn, HCl
(a) SnCl2 yields Sn2ϩ and ClϪ
(b) Disproportionation of Sn2ϩ yields Sn4ϩ and Sn
(c) SnCl2 and S yields Sn2S3 and SnCl4
H2S
H2
H2ϩ, Hϩ, S, S2,
S2Ϫ, S22
(d) Sn4ϩ and 2S2Ϫ yields SnS2
(e) Sn and Sm yields SnSx and SmϪx
(f) SnCl2, S and Hϩ2 yields SnS and HCl
H, H2ϩ
(g) SnS and S yields SnS2
(h) Sn2S3 and S yields SnS2
Downloaded on 2014-06-01 to IP 155.33.16.124 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

3716
Journal of The Electrochemical Society, 147 (10) 3708-3717 (2000)
S0013-4651(00)02-060-7 CCC: $7.00 © The Electrochemical Society, Inc.
form the Sn2S3 compound and the other 50% form SnCl4 again. The
overall reactions that could take place is given by
measured film thickness corresponds only to this compound. Thus, a
reduction in Rd from the previous value for g is expected. For 0.49 <
g < 0.6, the deposited material is mainly SnS2, and the measured film
thickness corresponds to this compound. Thus, the value of Rd in this
interval must be lower than the one found in the interval discussed
for the case of Region II, but would follow the observed increasing
trend due to the increase in the concentration of SnCl4.
2SnCl4 ϩ 3H2S ϩ 1 H2 r Sn2S3 ϩ 8HCl
2[SnCl4 ϩ H2 r SnCl2 ϩ 2HCl]
1 H2S r H2 ϩ (1/8) S8
(1/8) S8 ϩ 1 H2 r H2S
[12]
Region IV.—In this region there is the formation of a mixture of dif-
ferent compounds as shown in Fig. 8. The deposition rate shows a
small increasing trend due to the incorporation of several compounds.
which can continue until the S concentration goes down. At this
limit, the formation of SnCl2 is expected. This explains the incorpo-
ration of SnCl2 in the deposited material prepared for g ϭ 0.49 (see
Fig. 7b).
In the region of 0.2 < g < 0.5, the atomic concentration of sulfur
is higher than that of tin before starting the glow. During the glow,
the chemical reaction represented in Eq. 12 takes place, but due to
the fact that there are more S than Sn, part of the generated S are ad-
sorbed on the growing Sn2S3 thin film surface yielding the formation
of SnS2 through the reaction
All the data points shown in Fig. 5 are mean values with the cor-
responding standard deviation. If this is considered, then in region II,
the deposition rate for both compounds is practically independent of
the precursor concentrations. Solomon et al.33 reported that, for
Si1Ϫx Cx thin films prepared by the decomposition of a mixture of
SiH4 and CH4 by PECVD, the deposition rate is independent of the
silane concentration as a consequence of the low power regimen
conditions (silane is decomposed by the plasma while methane is
not). The conditions reported in this work are similar to that in
Ref. 33, thus, a similar behavior is expected. Nevertheless, the chem-
ical reactions are not the same for different precursors, and hence,
distinct results may be obtained. This is evident in the case of other
g regions for which the main compound deposited is SnS2. In those
regions, a linear relationship between the deposition rate and the
SnCl4 concentrations may be obtained. Further work in this direction
will be carried out in order to confirm this observation.
Sn2S3 ϩ S r 2SnS2
[13]
Thus, since not all of the Sn2S3 material is transformed into SnS2,
the Sn2S3-SnS2 mixture may be produced. This explanation supports
our results obtained for 0.2 < g < 0.49 for which the deposited mate-
rial presents both compounds.
In the case of 0.6 < g < 1, the reaction chamber has more mole-
cules of Sn than S. Then the following reactions could take place.
1. Reduction of SnCl4 by H2 giving SnCl2 and HCl.
2. Decomposition of H2S into H2 and Sm through a reaction such
as Eq. 10.
Conclusions
This work presents a systematic study on the deposition and
structural properties of SnxSy thin film materials prepared by
PECVD using SnCl4 and H2S as precursor materials and H2 as dilu-
ent. It was found that a “low-pressure” regime for which the plasma
decomposes H2S does not decompose SnCl4. The chemical reactions
used to explain the film formation were separated from the compli-
cated chemistry of plasma owing to the difference in the decompo-
sition thresholds between both precursors. In this regime g deter-
mines the relative chemical composition, the crystallinity, and the
preferential growth of the deposited SnxSy thin films. These films are
of polycrystalline nature for all the g values considered here. For g <
0.2, EDS measurements have shown that the deposited material is tin
disulfide with a stoichiometry close to 1:2. XRD measurements
show that this compound presents only the 2H-SnS2 phase with a
hexagonal structure showing a preferential growth along the [001]
direction, and hence, the c axis is perpendicular to the plane of the
substrate. All these films present a very smooth surface and high
structural quality. For g > 0.2, XRD measurements show that the
deposited material is a mixture of SnS2 and Sn2S3 compounds. It
was found that for g ϭ 0.49 the deposited material is mainly Sn2S3.
The precursor concentrations do not affect the crystalline grain size
of the SnS2 compound. However, the deposition rate depends on the
g value, because the deposited film is composed of different materi-
als and phases. The maximum deposition rate found for the SnS2
compound was 8 nm/min, while for films having the mixture of
SnS2-Sn2S3, it is of the order of 35 nm/min. On the other hand, it
was found that the prepared thin films with g > 0.1 have chlorine
incorporation, even though small. The hydrogen flow rate added dur-
ing the process has a gettering type of action, reacting with the chlo-
rine related species generated. However, this effect was not enough
to avoid the Cl incorporation at the deposited thin films. More work
is necessary to establish the deposition parameters for producing
thin films without chlorine. Here, we found specific deposition con-
ditions that lead the SnS2 thin film grows with a columnar oriented
growth along the c axes with a deposition rate of 8 nm/min. The
crystal grain size of such films is 13 nm. Due to the fact that the SnS2
thin films prepared under this condition had shown an n-type elec-
trical conductivity, this work opens up the oportunity to use this
material for building an n-SnS2/p-SnS heterojunction prepared com-
pletely by PECVD.
3. A reaction of SnCl2 with Sm to form SnS2 and by-products like
SnCl2, HCl, H2, and SmϪ2
.
All these reactions take place until g goes to 1. Therefore, it is
possible to obtain a solid thin film material formed by several com-
pounds, mainly SnS2 and SnCl2 for 0.6 < g < 0.75; and SnS2, S,
SnCl2, and even metallic Sn for g > 0.75, as can be seen in Fig. 8a
and b.
On the other hand, during the plasma process, tin ions (Sn4ϩ
,
Sn2ϩ) and atomic Sn, produced by the reduction of SnCl4, coexist at
the same time with ions of sulfur (S2Ϫ), and it is expected that SnS
compound could be generated through Reaction f in Table IV. Nev-
ertheless, XRD did not show evidence of this material in the deposit-
ed thin films. This observation does not exclude the formation of
SnS. Perhaps the excess of S present will react with SnS to form
SnS2, in the same way as with Sn2S3.13 Thus, the stoichiometry of
the deposited film is shifted toward SnS2 as has been shown in the
present study.
Since there are several reactions that yield to different com-
pounds in the range of g chosen in the present study, the dynamics
of the growth process strongly depends on the concentrations of the
precursors. Their effect would be reflected in Rd as well. Based on
the results obtained and considering that the films are deposited by
an atomic (molecular) growth process (typical in PECVD), it is pos-
sible to explain the behavior of Rd shown in Fig. 5, as follows.
Region I.—When the SnCl4 concentration increases, the atomic con-
centration of Sn increases and more atoms of tin are available to
form the SnS2 compound; hence, Rd increases.
Region II.—Here two different materials are growing at the same
time. The jump in Rd showed at the transition point around g ϭ 0.2
can be explained by considering that the measured film thickness
corresponds to both SnS2 and Sn2S3. In this region, more precursors
of tin-bearing species are involved in the growth process because the
SnCl4 concentration increases. Thus a trend toward an increase in Rd
is expected, but due to the growth competition between both com-
pounds, the effect is not very pronounced.
Region III.—Here for specific values of g, a specific compound was
obtained. For g ϭ 0.49, the deposited material is Sn2S3 and the
Downloaded on 2014-06-01 to IP 155.33.16.124 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 147 (10) 3708-3717 (2000)
3717
S0013-4651(00)02-060-7 CCC: $7.00 © The Electrochemical Society, Inc.
13. R. D. Engelken, H. E. McCloud, C. Lee, M. Slayton, and H. Ghoreishi, J. Elec-
trochem. Soc., 134, 2696 (1987).
Acknowledgments
We wish to thank Leticia Baños (IIM-UNAM) for XRD analysis,
and Jose Guzman (IIM-UNAM) for SEM and EDS analyses. One of
the authors (A.S.-J.) is thankful to the PADEP-UNAM program for
the financial support given through project no. 5338 in the initial
stage of the work. CONACyT, Mexico, has supported this work
under contract 3723-PA.
14. M. T. S. Nair and P. K. Nair, Semicond. Sci. Technol., 6, 132 (1991).
15. S. López and A. Ortíz, Semicond. Sci. Technol., 9, 1 (1994).
16. S. López, S. Granados, and A. Ortíz, Semicond. Sci. Technol., 11, 433 (1996).
17. N. Koteswara Reddy and K. T. Ramakrishna Reddy, Thin Solid Films, 325, 4 (1998).
18. A. Ortíz, J. C. Alonso, M. Garcia, and J. Toriz, Semicond. Sci. Technol., 11, 243
(1996).
19. W. Albers, C. Hass, and F. van der Maesen, J. Phys. Chem. Solids, 15, 306 (1960).
20. J. J. Lofersky, J. Appl. Phys., 27, 77 (1956).
21. M. Parenteau and C. Carbone, Phys. Rev. B, 41, 5227 (1990).
22. R. Whitehouse and A. A. Balchin, J. Cryst. Growth, 47, 203 (1979).
23. G. Said and P. A. Lee, Phys. Status Solidi A, 15, 99 (1973).
24. M. A. Lieberman and A L. Lichtenberg, Principles of Plasma Discharges and
Materials Processing, John Wiley & Sons, Inc., New York (1994).
25. A. Von Engel, Ionized Gases, 2nd ed., Clarendon Press, Oxford, U.K. (1965).
26. CRC Handbook of Chemistry and Physics, 71st ed., D. R. Lide, Editor, CRC Press,
Boca Raton, FL (1991).
27. J. C. Alonso, R. Vazquez, A. Ortíz, and V. Pankov, J. Vac. Sci. Technol. A, 16, 1 (1998).
28. I. Traus and H. Suhr, Plasma Chem. Plasma Process., 12, 275 (1992).
29. H. P. Klug and L. E. Alexander, X-Ray Diffraction Procedures for Polycrystalline
and Amorphous Materials, John Wiley & Sons, Inc., New York (1954).
30. T. Shibata, Y. Muranushi, T. Miura, and T. Kishi, J. Mater. Sci., 26, 5107 (1991).
31. S. C. Ray, M. K. Karanjai, and D. DasGupta, Thin Solid Films, 350, 72 (1999).
32. C. Wang and R. G. Cheng, Philos. Mag. B, 53, 183 (1986).
33. I. Solomon, M. P. Schmidt, and H. Tran-Quoc, Phys. Rev. B, 38, 9895 (1988).
34. Z. Zainal, M. Z. Hussein, and A. Ghazali, Sol. Energy Mater. Sol. Cells, 40, 347
(1996).
CONACyt (project 3723-PA) Mexico, and Universidad Nacional Autono-
ma de Mexico assisted in meeting the publication costs of this article.
References
1. E. Mooser and W. B. Pearson, Phys. Rev., 101, 492 (1956).
2. G. Busch, C. Frohlinch, and F. Hulliger, Helv. Phys. Acta, 34, 359 (1961).
3. W. Albers, C. Hass, H. J. Vink, and J. D. Wasscher, J. Appl. Phys. Suppl., 32, 2220
(1961).
4. G. Valiukonics, D. A. Guseinova, G. Krivaite, and A. Sileica, Phys. Status Solidi B,
135, 299 (1990).
5. M. T. S. Nair and P. K. Nair, J. Phys. D: Appl. Phys., 24, 83 (1991).
6. N. Koteswara Reddy and K. T. Ramakrishna Reddy, in Proceedings of the IEEE
26th Photovoltaic Specialist Conference, 515 (1997).
7. G. Domingo, R. S. Itoga, and C. R. Kannewurf, Phys. Rev., 143 (2), 536 (1997).
8. E. Schonerr and W. Stetter, J. Cryst. Growth, 30, 96 (1975).
9. T. Chattopadhyay, J. Pannetier, and H. G. von Schnering, J. Phys. Chem. Solids, 44,
879 (1986).
35. B. Chapman, Glow Discharge Processes: Sputtering and Plasma Etching, John
Wiley & Sons, Inc., New York (1980).
36. I. Trauss, H. Suhr, J. E. Harry, D. R. Evans, and E. Andrade, Plasma Chem. Plas-
ma Process., 13, 77 (1993).
10. J. George and C. K. Valsala Kumari, J. Cryst. Growth, 63, 233 (1983).
11. K. Kourtakis, J. DiCarlo, R. Kershaw, K. Dwight, and A. Wold, J. Solid State
Chem., 76, 186 (1988).
12. S. K. Arora, D. H. Patel, and M. K. Agarwal, J Mater. Sci., 29, 3979 (1994).
Downloaded on 2014-06-01 to IP 155.33.16.124 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).